The reality of Enhanced Campaigns today doesn’t create much urgency to switch over. We don’t get any more controls, and in fact, we lose some. However, the near-term future of EC is very exciting, particularly with respect to improved location targeting.

Chain business operations from retailers, to package delivery firms, to insurance companies with local agents will see the true benefits of the new structure; hopefully, soon.

Enterprise Programs Need More Data

Google has a few thorny issues to address to unlock the full potential of advanced geo-targeting using the Enhanced Campaign architecture. They have to reveal enough information to platform providers and technology shops like RKG so we can bid smartly.

At the same time, they have to respect the privacy agreements with users, particularly users of mobile devices. Users may give Google permission to know their precise location, but don’t really give it permission to share that information with others.

Enterprise advertisers also don’t really need to know a user’s longitude, latitude and altitude, either. What we do need is for the information Google shares with us to match the bidding and targeting controls we have through Enhanced Campaigns. When advertisers can carefully measure the performance variations associated with each factor, we can react to them wisely.

Google can provide us with fabulous controls without revealing too much about users by passing us two (ideally, three or more) pieces of information through valuetrack parameters (or via API tied to a click id that matches the GCLID passed):

1. Postal code. Whether through postal code, city, state or country, this information combined with census data can be used effectively to understand how regions and types of regions impact performance.

We can currently map IP addresses to zip codes, but that information is pretty far from ideally accurate; and, the mapping of IPS to zips is a pain. Google has click-level city, state data available through the API, but the click id there isn’t the GCLID passed in the redirect; so, we can’t yet connect it to the conversion and post-conversion information we have. We’ve asked them to address that.

The power of being able to model data from different regions looking at all types of attributes of that region, and then be able to implement bid adjustments to easily act on that modeled data is exciting.

Carving up programs for geo-targeting is a painful process and we end up not being able to get very granular in our approach to this because of campaign limitations. The more a company spends in paid search, the more money is at stake in this level of granularity.

2. Relative location. For a brick and mortar business, we don’t need to know exact location, as long as we know where someone is in relationship to us. Knowing business addresses, Google could give us the distance a user is from our nearest store/office without revealing the person’s precise location.

This would provide crucial information for understanding how proximity influences advertising value.

3. Context of location. Is the user stationary, walking, or riding? Is the user at “home,” “office,” or “other”? Again, Google could give advertisers cues to the user’s intent that would likely impact ad value, and therefore, allow us to target smartly. Passing each context in a separate valuetrack parameter would be awesome.

Benefits Of Laser Targeting

Conspiracy theorists might suggest that Google really doesn’t want advertisers to have all this information or to bid smartly, at all. I’ve heard folks suggest that this is really nothing more than a mechanism to make it more difficult for advertisers to target consumers. I don’t buy it because that would actually hurt Google in the long run.

Google should not fear laser targeting. Folks in Mountainview might think, “if advertisers can really target the best users, they’ll spend less on other users and that might hurt our revenue.” Not so. “Cherry picking” actually benefits Google and advertisers and users, for that matter. There are several reasons for this:

• Right-sizing the competition. To the extent that there are “good” targets and “bad” targets, advertisers will spend less to attract the latter, but more to attract the former. Differential bidding works in both directions and the net effect is a win for Google.
• My cherry is sometimes your lemon. Online pure-plays will pay more for those likely to convert online; brick and mortar businesses might pay more for users who do not convert online. Allowing each business to target their own cherries will lead to identification of more cherries.
• More bang for the buck. Smarter bidding leads to more efficient resource allocation; more efficient resource allocation leads to more bang for the advertising buck; more bang for the buck leads to more bucks. Advertisers that spend to efficiency objectives will spend more; those that spend to a budget will enjoy better ROI, which leads to a larger budget.

More efficiency helps advertisers get more from the channel cost effectively. This happens because users get advertising that is more relevant and useful to them, and it has the net impact of Google making more money — a three way win.

Cost-Effectiveness Of Laser Targeting

The notion of one bid plus a string of modifiers is not the ideal, but it is what we have to work with. The construct assumes that the contextual cues are independent of one another, which likely isn’t the case.

An advertiser might want to bid more for smart phone users within a mile of a brick and mortar location but not after the stores are closed. But, damping the bids after 5PM for smartphones also damps the bids for stationary tablets and desktops in that zone, which may not be what you want.

The ideal system would allow advertisers to pass Google a discrete bid for each combination of contextual cues, but that gets to be a great many discrete bids in a big hurry and would shatter the stated goal of simplicity.

Understanding the location, the relative position, and the user’s context will help advertisers, with the largest, most sophisticated enterprise firms benefiting the most. The benefits will flow to users in more targeted, and therefore, more useful ads; to Google, in more revenue; and to the advertisers who can take advantage in growing their program, in profitability. Moreover, Google can share all three without breaching trust with its users or sharing personally identifiable information.

Who’ll join the call for more data?

Much of the early commentary has focused on the version 1.0 implementation. It is important for Enterprise SEM program managers to start thinking hard and deep about the implications of Enhanced Campaigns in the long term, as well as understanding what it is likely to become.

Click here to view the embedded video.

Better Architecture To Handle More Complexity Is Not Simplification

Google has mistakenly framed this as simplifying targeting. In fact, it’s just a different mechanism to accomplish the same goal.

In the past, if you wanted to have different bids, different landing pages, and different ad copy based on device type, you had to create replicated campaigns. Now, you just need one campaign, but you STILL have to provide all the same targeting information to Google with respect to what copy goes with what device, which URL to use, etc.

The problem with the old structure was that the number of replicated campaigns required to target by device, geography, syndication preference, etc., compounded geometrically. This was awkward from a management perspective, and likely caused inefficiencies in Google’s auction algorithms, as well. The single campaign with multiple levers approach is cleaner in that sense, but no less complex to do well.

As I described earlier, this change will actually add to the complexity of bidding in that instead of setting a bid for an ad that is associated with a keyword, match type and context, you have to set a bid for a keyword and then set multiplication factors for each different context.

This may make it simpler for the SMB willing to give Google their bank routing number and freedom to serve ads as Google sees fit, but I don’t see any Enterprise advertisers choosing that route. Two levers instead of one adds to the complexity of bidding.

It Is All About Context

Larry Kim has it right, it’s all about context. Many folks are upset with Google because the distinction between traditional computing devices (desktop and laptop) and tablets is gone in Enhanced Campaigns v 1.0. That is an error on Google’s part, and I’ll bet they fix it pretty soon.

Google’s research shows that the behavior of people using tablets in the context of being at home in the evening is indistinguishable from their use of computers in that context. Google has more data on this than anyone (possibly the understatement of the year), so I don’t doubt that’s right. But this is just hinting at the real future of enhanced campaigns.

The controls in v 1.0 are rudimentary (geo, device, time of day/day of week). But, let’s think about the future controls and all the amazing insights Google has about users that could be subject to targeting.

From your online behavior Google knows:

• your demographic profile
• what you do and say socially
• what websites you like and/or frequent
• your religion
• whether you transact business online or prefer to do business in brick and mortar locations
• what you’ve searched for ever, what you’ve searched for recently
• whether you’re just researching or seriously comparing products/services and offers
• what type of device you’re using
• the size of the screen
• the connection speed

And the list goes on and on…

Now, let’s add to that what Google knows from your device GPS, its own knowledge of geography, and your login to Google services:

• where you live
• where you work
• how you commute
• how fast you drive
• where and when you go to lunch
• whether you’re walking or riding in a car or riding a subway
• They may be able to tell whether you’re sitting or standing
• They could probably know who’s cheating on their spouse (look for two devices that are usually together at the same residential address… look for one of those devices spending the night at a different residential address when the other device is out of town at a hotel….)

…it is truly mind-boggling!

Campaign replication simply could not keep pace with this. The architecture had to change. You can bet your life that some or all of these contextual cues will impact the type of ad that will resonate most with that user at that moment, and how each potential advertiser should value that ad impression.

Used in the context of display advertising, this kind of information can be horribly creepy and bad. “Looking for a divorce attorney? You should be!” However, there is no creep factor with search ads. Not only do we understand a great deal about this user in this moment, we also know that they’re interested in seeing what we have to offer right now. The act of searching creates a “permission marketing” experience that makes ads relevant and useful in a way that no other type of advertising can since the print yellow pages  offer.

How will we know what ads to serve and what bids make sense in different contexts? Google will have to give us contextual information about each user who interacts with our ads. This will be done by passing a click ID that can be used as a key to pull context from Google. That, in turn, will allow us to see which contextual elements and combinations of elements impact post click performance. It will allow us to test different creative and landing page experiences so that we may learn what works best in a given context.

Keywords will remain King. I wager that we will all find out that the words used in the search are the most important predictor of value to an advertiser, but it would be crazy to think that those other contextual factors won’t be important as well. If context isn’t King, it will certainly be Duke.

Google isn’t sharing this data with anyone yet, and they won’t be willing or able to share everything they know about every user for privacy reasons. However, advanced advertisers should be furiously working on how they will gather and process the information Google passes and how they will react to it. The mechanics of what we do will have to change, but there are fundamentally new and exciting possibilities that smart marketers should start grappling with now.

Image via Shutterstock

The smartest way to begin 2013 might well be to make a careful review of paid search performance and practices in 2012.

If your company spends closer to 8 figures than 6 on paid search annually it is well worth the time of the CMO to take a deep dive here. Evaluating an enterprise paid-search program can seem a daunting task.

Complexity and scale can lead senior marketing leadership to make one of four mistakes:

• Allow paid search managers to evaluate themselves

• Allow a vendor seeking to win your business to evaluate your team’s performance

• Hire a consultant who knows little about paid search to do the evaluation

• Worst of all, evaluate performance against aspirational goals set by others

For obvious reasons, none of these yield a fair assessment of how paid search performed against opportunity.

Instead, CMOs and marketing directors should take the bull by the horns. The answers to 28 questions in 3 key areas can allow the marketing leadership to gain a better understanding of paid search, better understand the challenges and opportunities the channel presents, and get a clear picture of how well her/his team performed against that competitive landscape.

Stage 1: Goal Review

Before reviewing any data, it is important to get a clear understanding of what paid search is expected to do for your organization. This isn’t a question of what the budget or forecast was, it is much more fundamental than that.

It is often remarkable how little thought goes into these questions. Deeply probing what we’re trying to accomplish and how we measure our performance against those objectives is a critical first step.

Image via Shutterstock used under license.

1. What do we seek to achieve through paid search advertising? Is the goal to make money in the immediate term, short term, or long term? Is it a branding exercise? Some hybrid?
2. Have we included in our goal setting some notion of long term value of a customer? Should we?
3. Do we correctly separate performance of our brand keywords from competitive non-brand keywords? {This is a pass/fail question. Blended performance goals make no sense.}
4. Have those goals changed over time? Since last year? During the course of the year? Why did the goals change?
5. What metrics do we use to know whether we are achieving our goal? Is it a single success metric? If so, is that metric as closely tied to value as it can be?
   • In e-commerce: why use orders as a metric instead of sales, and why use sales instead of margin?
   • If leads are the success metric, do we assess the value of those leads? Do the values vary by keyword, geography or device? Do we use those differences in our bidding?
6. How do we measure offline success driven by paid search? If we cannot measure directly, why can we not estimate it?
7. Are there other success metrics we should use? Should not a “get directions” click or “click-to-call” on a smartphone count as some type of success?
8. How do we address the issue of multiple paid-search clicks preceding a successful visit? Do we parse credit appropriately?
9. As a marketing team, how do we think about multiple marketing interactions preceding a successful visit (online or off)? Do we parse credit between channels? Should we?
10. Do all of our business units follow the same practices? Is paid search managed by different teams within the organization? Does that make sense?
11. If applicable, how do we manage international search? Do we have language skills internally to do this well? Do we advertise on Yandex (Russia) or Baidu (China)? How do we measure the success of those efforts?

Stage 2: Performance Review

How did we do against the goals we pursued? Whether the goals and metrics were the right ones, whether they changed during the course of the year rightly or wrongly, the question here should be how did the paid-search program do against the goals they sought? Performance shouldn’t be faulted for achieving the wrong goal.

Again, we should look at this not as a function of performance against a forecast or budget, but against opportunity. Results can be above or below forecast for reasons relating to forecasting methods and foibles, not performance. The numbers and responses in this section, and the responses in Stage 3 will get us what we want.

1. Show me aggregated paid search performance data by month, splitting out brand and non-brand.* If there are apparent deviations between the performance and the goals month-to-month, what is the explanation for those deviations? If non-brand efficiency metrics vary significantly, why would we as an organization be willing to invest more in marketing some times than we are other times?* If the goal was to achieve an aggregated performance efficiency, we can’t fault the performance of non-brand advertising in isolation, but we should at least see it clearly and understand what it means. The non-brand performance is the only piece over which the paid-search team has meaningful control, so that data is what must be used to evaluate performance.
2. Show me the Year-Over-Year non-brand performance by month. What factors drove the changes year to year?
3. Show me the trend in the fraction of total website success metrics driven by paid search. Split this out by brand and non-brand, as well. Why is it trending up or down? What does that mean for the business? Does that say anything about performance?
4. Show me non-brand performance data by month, split out by search engine. If there are differences in marketing efficiency, what is the explanation?
5. Show me Google-only performance over 2012QX {Marketing leader picks the X} broken out by category (campaign might be a decent proxy). Do the efficiency differences make sense?
6. For a different quarter, show me Google performance data by Keyword. Sort this data by advertising cost descending. Understanding that there is statistical noise involved, does the KW level performance look reasonably coherent? Do major differences in performance have a sensible explanation?
7. Bucket this list of keywords by ranges of click volume so that there are roughly an equal number of clicks in each of 5 buckets ranging from the highest traffic terms to the lowest traffic terms. Do any performance differences between these buckets have a reasonable explanation?

There are reasonable explanations for many types of anomalies. The point of this exercise is to gain a deeper understanding of the program’s performance against goals, and to see how the paid-search team answers hard questions about performance. If the answers make sense, great. If the answers smell fishy, they probably are.

Stage 3: Practice Review

Even if the performance numbers, trends and answers to the questions in Stage 2 look good and sound reasonable, it is impossible to tell whether the program is truly optimized without scrutinizing the team’s practices as well. Efficient numbers could still reflect a poorly-built program covered up with decent bid management.

1. How many distinct keywords are active in our Google account? Our Bing Account? Why are those different numbers? How many distinct keywords have we tried? Why have we not tried more? Why do we turn off keywords? Couldn’t a case be made that there are no bad keywords, only bad bids? Looking through the keyword-level data in #6 above, you might notice some holes; ask about them specifically and expect a reasoned response.
2. What do we do to reduce poor quality traffic? The answer should include references to match type layers, negatives, and there syndication partner treatments? Go into the account with the paid search manager and drill into a campaign and adgroups. Are there obvious negative phrases that are missing? Is the ad copy associated with each adgroup compelling but also accurately reflective of your brand? Do you see different match types running for the same keyword? If so, is more being bid for the exact match traffic than broad match? If so, that’s probably a good sign; if not, it’s probably a bad sign, but another good avenue to question.
3. While you’re in the account, navigate to a different campaign and look through the ad copy and landing page assignments. Do the keyword, copy and landing pages fit together? The landing page should reflect the specificity of the user’s query — no deeper, no shallower.
4. Show me some examples of copy tests we’ve run in the last year. What lessons did we draw from them? What copy tests had the biggest impact on performance? Do we change copy without testing? Has that proved to be valuable? What fraction of your time goes into ad copy versus other tasks? Does that balance make sense?
5. Do we split out campaigns by device? {Best case three separate campaigns for desktop, tablet and smartphone. Acceptable case and often just as good, two campaigns: desktop + tablet in one and smartphone in the other. Worst case is two campaigns desktop in one, tablet + smartphone in the other.} Do we have different ad copy for different devices? Do we have different objectives for different devices? Should we?
6. Do we split out campaigns by geography? Should we?
7. What other types of segmentation do we do to better target bids and copy?
8. How do we set bids? {If the response involves ‘trying to find the right position,’ it’s time to find a new manager.} How often do we adjust bids? Are we adjusting based on time-of-day and day-of-week? Do we have mechanisms in place to anticipate seasonal flux? Do we make use of bid simulator data to understand the cost/benefit tradeoffs of different bids?
9. Do we make use of all the available advertising vehicles and options associated with search these days? Are we using:
   • Product Listing Ads (mostly an e-commerce option)?
   • Dynamic Search Ads?
   • Sitelinks?
   • Seller Ratings (if applicable)?
   • Search Retargeting? {This one is new, but really important.}
10. How do you prioritize activities? Do you spend most of your time on the highest leverage activities? If not, what keeps you from doing that? The answer might be that task lists from on high or demands from other constituencies in the company prevent the team from using its human resources wisely. If so, that’s important to understand, and address.

Conclusion

This review process does not require the reviewer to have deep knowledge of paid search, the reviewer simply needs to have common sense, and a good nose for BS. Paid search makes sense when done well. Data that doesn’t make sense combined with answers that don’t make sense should raise big, red flags about the health of the program.

If the practices sound right, the data makes sense given the goals and constraints placed on the team, and the team has smart, rational make-sense answers for anomalies, then the team is doing its job well against the opportunity available. The right goals combined with the right practices, execution and technology will produce the best data possible.

May 2013 bring you the best results from paid search ever!

The demands on the online marketing director’s time have never been greater or more fractured, and it only gets worse from here.

The greater the array of programs to manage, the more essential it becomes to prioritize. There is a natural human tendency to focus attention on the newest, shiniest object in the room, regardless of its importance in the great scheme of a business, but losing sight of the core can have devastating consequences.

Every person who oversees online marketing should, therefore, make the following 3 resolutions:

Shutterstock image used under license

1. Prioritize Budget Effectively

I resolve to prioritize attention based on the size of the marketing budget for each channel and its importance in driving incremental revenue, not based on the novelty or coolness of the program.

We see, time and again, prospects who want to talk to us about sophisticated attribution methodology when their paid search program is a shambles, their site can’t be crawled, and their display advertising budget is being spent without proper tracking or ROI metrics in place.

We’re big believers in the value of fine tuning and testing the latest and greatest, but as exciting as multivariate testing in Facebook advertising sounds, that should draw your attention only after your search marketing efforts are thoroughly buttoned-down. The return on effort will be much greater when applied to areas where the most opportunity lies.

Easier said than done!

The difficulty in this usually takes one of two forms (sometimes both):

1. Corner Office Interjection: “I read an article about X, what are we doing about X?!? X should be your top priority.” “Our arch-rival competitor is doing Y, what’s our Y strategy? Stop wasting time on search and email and focus on Y!”
2. Self-Awareness Limitations: Oftentimes, advertisers think their core programs are in better shape than they are. “We have a top rated agency managing this, so we’re good.” “Our in-house SEM manager had many years of experience prior to joining us, so we’re good.”

Resolving difficulty #1 is tough. It requires education… which is difficult… and, when education fails, enough gumption/job security to tell the big boss that s/he isn’t in position to evaluate marketing priorities. In other words: this problem is often insuperable.

The answer to #2 is simply: “Trust, but verify.” We should never put faith in ratings, rankings, reputation, or “experience.” We should put faith in strategy, practices, execution and results. Every program from paid search to SEO to email, CSE, display, affiliate, etc. needs to be evaluated regularly based on changes in the landscape.

The world of online marketing is in constant flux, and staying on top of the latest opportunities within each channel is no small task. Excellent managers welcome performance evaluations.

Shutterstock image used under license

2. Look Below The Surface

Within Paid Search, I will pay attention to what matters most, not simply what is most obvious on the surface.

Because it is most apparent, ad copy often garners the most attention and the great preponderance of management time. That’s rarely a wise allocation of effort. There is terrific ROI on effort in writing targeted and compelling ad copy. There is sometimes value in turning copy with offers; but, the value can be illusory, improving CTR but damaging value-per-click to make it a wash.

Copy changes effectively reset the Quality Score to ‘average’ until the copy proves itself to be better or worse, meaning frequent flips hurt Quality Score and prevent capturing the benefits of strong control copy.

The odds that the time spent flipping copy to catch a 10% discount on tube socks is worth the cost of that time is very very small.

The right reasons to change ad copy include:

• Demonstrably improving results against the control
• Updating brand messaging

The wrong reasons to change ad copy include:

• You’re bored with the existing copy
• You read somewhere that constantly testing copy was the best use of your time
• You don’t know what else to do
• The boss likes to see stuff happening

Outsiders don’t understand the importance of research, analysis, and the impact that these projects have on targeted bidding. Outsiders believe that an automated bidding system eliminates the need for smart human analysis and tuning. People who know their stuff know better, but sometimes can’t focus on the levers that matter because they’re forced to chase the tip of the iceberg, instead of drilling into the core where the most impact can be had.

Management By Objectives applied poorly, often results in a search manager’s performance being judged by the quantity of work done rather than the quality and value of the work. As John Wooden once said: “Never mistake activity for achievement.” It is trivially easy to keep an army of staff busy working on an enterprise paid search program; but, absent knowledge, all that work can be counter-productive.

Paid search is part art, part science. Neither artists nor scientists produce their best work when used as “a pair of hands” by non-artists or non-scientists. Creativity and ideation in paid search require time set aside for thought. Productive thought can look like idleness to the uninitiated. The way to tell the difference is to look at the ultimate results.

Shutterstock image used under license

3. Evaluate Success

I will evaluate success based on strategy, execution and results against opportunity, not based on performance against wishful thinking forecasts.

Enterprise paid search managers cringe when they hear the big boss announce: “We have very ambitious goals for paid search next quarter!”

Particularly when, as is often the case, the paid search manager did not participate in the forecasting process.

Budgets and forecasts are necessary even if they are inherently impossible, but when they are developed to fit company aspirations rather than carefully constructed based on anticipated market opportunity, they can become downright detrimental.

“We plan to double the paid search program next year at an even better ROI!”

To a top-drawer paid search manager with a program hitting on all cylinders this sounds a great deal like: “Get your resume together. We don’t understand what you do, and plan to make your year miserable if you stick around.”

We face this as an agency on occasion, and sometimes it is a product of our own past success. Taking over management of a badly broken program and generating a 300% growth in revenue on improved ROI does happen, but it can lead to expectations that those kinds of growth rates will happen every year.

Again, this reflects a failure on the part of the paid search manager to successfully educate the people who create these budgets and to force themselves into the budgeting process. Nevertheless, when a budget is presented fait accompli, it can be very difficult to focus attention on what is and what is not within the realm of possibility.

Even when the paid search manager participates in the forecast, forecasts should not be the basis of evaluation. Accurately forecasting paid search performance would require knowing not only what consumer behavior will be 3, 6, and 9 months down the road.

We can’t serve ads if people don’t search for them — it would require knowing what changes the engines will make to the page layouts, what changes they will make to their matching algorithms and partner networks, and what new controls and ad formats might role out impacting performance.

We’d also have to know what our competitors’ strategy and execution will be in the coming year, or whether new competitors will enter the marketplace. We don’t know and can’t always predict what will happen in our space any more than a stock analyst can tell you today what stock you should buy 6 months from now.

Forecasts serve a purpose for staffing, merchandizing, and accounting, but serving those needs demands frequent re-forecasting to adjust to present realities. Performance should be evaluated by how well the paid search marketer performed given the opportunity landscape they faced, not by how well they were able to predict what the results would be.

If you’re an online marketing director, please make these pledges for the New Year. If you’re a lowly paid search manager, consider slipping this under her/his door!

Have a great 2013!

Images from Shutterstock, used under license.

• Senior executives reviewing data from programs they don’t understand
• A shortage of institutional memory
• Consultants coming and going who offer opinions based on superficial data

Each of these traits make performance reporting particularly important, and sometimes challenging, for paid search. These challenges are often most keenly felt when the enterprise decides to make significant changes to the goals of the program.

Any well-managed paid search program* is subject to the law of diminishing marginal returns. The last dollar spent has a lower ROI than the first dollar spent because we buy the most cost-effective advertising first. Fans of RKG Blog are probably tired of hearing about it!

*Note that a badly managed program may not be subject to the same law. Improving volume and efficiency simultaneously can happen when poorly managed programs are placed in the hands of experts with great tools.

What this implies is that there is an inherent tradeoff between volume and efficiency at any point in time. Making the program bigger today will also make it less efficient, all other factors being equal.

Enterprises Change Goals

Business needs, changes in management, competitive pressures, or any combination, can lead to significant changes in paid search goals. Where the enterprise was using paid search for brand awareness and customer acquisition and was happy to invest in future growth, now the goal shifts to making search profitable in the immediate term, or vice-versa.

The beauty of paid search is that we can turn on a dime to hit those new goals.

The problem is that flawless execution will produce “bad” year-over-year comparisons that are a natural consequence of the change in goal. Senior executives in other departments and consultants (having no institutional memory) will proceed to beat up the paid search manager for doing her/his job well.

Intrepid SEM Manager Pat: “Last week paid search drove $1.5M in revenue at a 5:1 ROI”

Angry Memory-Challenged Executive Shawn: “YOY, paid search was down 15% last week, why do you suck?”

Pat: “Last year the ROI target was 2.5:1, so we had to pull back to achieve the 5:1.”

Shawn: “I’m not interested in excuses, I want to know why every week we’re down in paid search?”

Pat: “We’re doing really well! We’re twice as efficient, we just can’t match the sales volume at that efficiency.”

Shawn: “You call being down 15% doing well’?!? Shape up or ship out!”

Managers want to see graphs that go up to the right; they don’t want to see graphs that go down to the right (positive slope = good; negative slope = bad). In YOY comps, negative numbers in red are bad. The fact that those negatives were caused by other metrics turning in a positive direction is irrelevant to Shawn.

In fairness, these folks can’t have expert understanding of every number that crosses their desks. Perceived declines require explanation. That’s just managers being managers. The confusion isn’t the fault of Shawn; it’s our fault for not providing the right metrics on which to judge success in the new world order.

Problem: Excellent Performance Looks Bad

The real problem above is that the goals of the program shifted dramatically, but the metrics used to gauge its success did not.

In the case above, the enterprise clearly decided to move from a top line emphasis on driving revenue to a bottom line emphasis on driving profits, but the organization kept looking at the top line revenue numbers. Declines there looked like failure.

Let’s take a look at a sample data set to highlight the effect.

The above data from a fictional e-commerce website spending to a 2.5:1 ROI target for a year and 19 weeks, then shifting to a 5:1 target. For sake of argument, I assumed this would roughly cut the spend by 50% and the sales by 25% (that Law of Diminishing Returns notion).

For executives watching YOY performance of Revenue and ROI, they’ll certainly see the ROI improvement, but a decade of experience in the space suggests that they’ll nevertheless focus on the YOY revenue decline. “Wow, we were up 20 – 25% YOY, now we’re down?!?”

The YOY revenue data will look particularly bad.

Each week, the message will likely be: “Good job on the ROI, but what are you going to do about the revenue? We really need to get those numbers back up.”

Experience suggests that that conversation will get old after a fairly short time. Moreover, it causes the execs to second guess their own decision making in ways that don’t necessarily make sense for the business.

The Solution: Change The Metrics

Presumably, the change in goals reflects a realization that the program had been losing money, and a decision that it now needs to make money. Whether that is the right business decision or not, if the objective of paid search is to be more profitable, then profitability is the metric we should track.

A good metric to use here is Marketing Income. In e-commerce Marketing Income is defined as revenue minus cost of goods minus variable costs associated with credit card fees, pick-pack, etc., minus advertising costs.

If we assume an average margin of 40% for this company (obviously better if we can get the actual margin down to the KW level, but we’ll take what we can get), we have the following:

The wise SEM manager will say when the goal change is announced: “Terrific, we’ve decided to focus on profits. Let’s recast our reporting to show Marketing Income by week, and YOY Marketing Income improvement.”

Recasting the exact same data shows the following:

Note that instead of sending mixed signals with one metric going up (ROI) and one going down (Revenue), both metrics make a sharp turn in the desired direction. This isn’t an effort to hide the facts, it’s an effort to focus attention on the metrics we’re supposed to be improving and showing that, in fact, they’re improving.

But What About Lead Gen?

Suppose instead of e-commerce, we’re talking about a site that uses lead capture as it’s best metric of success. Let’s say this lead gen company has been willing to spend $25/lead, and now realizes that it needs to cut that figure to $15/lead.

If they’ve historically averaged 10,000 leads per week, that figure will obviously drop with more restrictive efficiency targets.

Yes, the CPL line in red drops, which is good, but so does that Leads per Week figure, and that will cause consternation unless the savvy search manager can change the conversation to better reflect the positive performance improvement.

The exact same logic we used with revenue to marketing income applies if we make a couple of assumptions. We may not have dollar figures and margin to reference, but if we make the simple assumption that the change in goals was precipitated by a realization that the leads weren’t worth $25 a piece in the past, we can plug in a value of what we do think they’re worth and get to a marketing income graph.

Are Senior Managers So Daft That This Switch Will Work?

Daft isn’t the right word. Busy, preoccupied, and unable to understand the details of every program is the more accurate. And, yes, controlling the discussion and focusing on the right numbers makes a huge difference in perception.

Is This Deception?

Are we hiding the truth from senior management when we do this? Absolutely not. We are focusing attention on the true goals of the program as they are today.

Driving great results is one part of the paid search managers job. Communicating those great results well to upper management is often every bit as important. Speaking truth to power is critical. Doing so in a way that accurately reflects the results of your work is wise, particularly if you do good work.

In most paid search reporting platforms, the default setting (often the only setting) creates a disconnect between conversion events and the marketing touches that drove them. Impressions, clicks, costs are tied to the day on which they occurred.

Conversion events are tied to the day on which they occurred. But interested customers don’t always convert on the first visit, or even the first day after that visit, which means some fraction of the conversions on any given day was driven by marketing touches that occurred on earlier days.

Day-Parting

RKG has argued for years that an important element of correct day-parting calculation is to tie the conversion events to the time of the click, not the time of the conversion. Day-parting allows sophisticated advertisers to bid more for higher quality traffic and avoid overpaying for lower quality traffic by measuring the impact the day of week and time of day have on traffic value.

This can only be done correctly by associating the conversion with the correct click-through. Since you bid for clicks, the right way to think about this is: of the clicks taking between 9AM and 10AM, what fraction converted?

Multi-touch interactions within paid search and across channels add a layer of complexity, but rarely alter the conclusions of a carefully done analysis as those effects are generally small and normally distributed. Creating time zone targeted campaigns may or may not be worth the additional management costs, but data should drive that decision.

The dissociated view — how many clicks happen in between 9AM and 10AM and how many orders happened between 9AM and 10AM — creates a somewhat different picture. Here’s an example of the conversion rate by hour measured three different ways: last touch, first touch and the dissociated view.

Adding Up/Down Bars highlights areas in which the dissociated view would lead to materially underbidding (white bars) and overbidding (black bars)

Difficulty Reading Tests & New Launches

The lag effect can also make it difficult to read the results of new campaign launches. Let’s say for a given advertiser in financial services that half of the conversions happen within 24 hours of the click and that the overall 21 day distribution looks like this:

Further, let’s say that the advertiser is willing to spend $50 to attract a qualified lead, and let’s assume that the brilliant Paid Search manager has this program dialed into the target efficiency from day 1.

Even with this perfectly optimized launch the program will appear to be significantly underwater for the entire cookie window simply because of the lag between click and conversion.

Indeed, the dissociated view (tying conversions to the time of conversion) only begins to show the real ROI of the new campaign after 21 days. That’s fine, as long as the advertiser is aware of the lag and doesn’t react too quickly to the apparent under performance.

Difficulty Dealing With Major Events

More common in ecommerce than other verticals: a big event, whether promotional or seasonal, often changes the value of traffic, not just the quantity of it. Absent accurate, detailed, historical performance data, intra-day bidding reactions can be tricky because we can’t see the “all-in” conversion rate of the traffic in real time.

A Useful Analytic Approach To Find A Solution

An excellent “hack” solution to this is to understand what normal conversion rates appear to be over shorter windows of time, like a day, or even an hour. Determining what fraction of eventual conversions takes place in the first hour (or on the first visit), allows you to take a pretty good guess at the “eventual” conversion rate.

The thinking is that if an event is expected to create a change in traffic value, and the “one-hour” conversion rate is measured to be X% higher than the normal rate, we can assume that the conversion rate over the full attribution window will also be ~X% higher.

Essentially what we’re doing is assuming that the shape of the conversion curve over time will be the same as it has been historically, and extrapolating early performance to project the eventual performance.

This same technique can be useful in estimating lead valuations and establishing LTV calculations. In long sales-cycle B2B and B2C businesses, it may take a year to get a clear picture of the average lead value from a given pool of leads.

Similarly, many advertisers are willing to take a loss to acquire customers based on the promise of lifetime value. Advertisers might lose money to acquire the customer even after the first “sale” because they believe they will recoup that loss and make profit off of future business from the same customer. Marketers look at lifetime value metrics historically to gauge how much they can and should be willing to lose to attract a new customer.

But how do they know that the one-year and two-year value of customers historically will be predictive of how these new customers from new sources will behave? How do we know that those new sales leads will convert over the long haul at the same rate as others we’ve received through different channels?

Well, we don’t. But what we can do to get a reasonably good sense of the matter is look at the typical 1-month conversion rate of leads, and if the new leads seem to show a similar conversion rate after the first month since capture then it isn’t crazy to assume that they will turn out to be of similar quality in the long run.

If the two-year customer value of a new customer is typically $200, It may be that $40 of that typically comes in the first month after the new customer came on board. So, with the new channel, we can’t see the whole two year value for…um…two years, but if the one-month value is ~$40 we might be reasonably confident that they are customers of equal value to the historic trends.

This isn’t an exact science. The nature of the event can change the click to conversion pattern, too, perhaps encouraging a larger fraction of the eventual buyers to “act now”. It could be that a one-hour conversion rate increase of X% may lead to an eventual conversion rate increase of something less than X%.

Historical data can teach us what types of events may shift the curve by how much, and what types do not impact the click to conversion pattern materially.

Similarly, new leads may convert at a different rate than normal and you won’t know for sure until much later. However, guessing that the historical patterns will hold up is almost always a reasonable starting point, and ignoring the challenge posed by lag time can lead to disaster.

Golden Tablets

I’m not talking about Joseph Smith’s discovery, I’m talking about Apple’s. The traffic from iPads is more valuable than traffic from desktops for many in the eCommerce sector.

Our data suggests the traffic from tablets as a class to be worth ~15% more than traffic from desktops. Traffic from the Kindle Fire seems to be less valuable, but that makes up a pretty small fraction of the total, therefore doesn’t degrade the overall value of the traffic significantly.

Different Demographics

Quite likely, the differences above are in part demographic. Folks with iPads are early adopters of technology with significantly higher than average household incomes.

The disparity with Kindle Fire may represent how ads are displayed on a generally smaller screen, but also, perhaps a function of the price point attracting a less affluent user.

Different Usage Pattern

People use tablets differently than desktop computers. I have heard from folks in the know that something like 80% of tablet traffic takes place between 6PM and 9PM in the tablet’s time zone.

More than 95% of tablet traffic happens over a wireless signal rather than a cell tower connection. Tablets are not, by and large, mobile devices. The tablet should be seen as a couch companion. People mostly use their online functionality while watching TV.

More Valuable & Less Expensive

Not only is the traffic on tablets more valuable than average, it is also less expensive by volume than desktop search. That is to say: the same bids on tablets put you “higher” on the page than they do on desktop. This is a function of the fact that few search managers are separating campaigns by device at all, and many of the ones who do have left tablets bucketed with smartphones.

In either case, the blended average traffic value is leaving opportunity on the table, and splitting tablets into a single “mobile” bucket with phones may be worse than doing nothing at all. Grouping tablet traffic with desktops may be an acceptable compromise for busy SEMs.

These facts should send online marketing directors for major brands scrambling to figure out whether and how tablets are being split out from traffic from other channels.

Also, those brand advertisers who spend significant money on television ads might consider whether the look and feel of the normal ad landing pages is consistent with the messaging in the commercials. Might there be an opportunity for a separate set of landing pages tied to the themes of the commercials running at the moment?

The Smartphone Conundrum

Your mileage may vary a great deal, but for many advertisers, particularly those with a wide range of products and services to offer, the online conversion rate for smartphones is a small fraction of that of desktop devices.

We have heard that this will all change with better, easier online payment types, and that NFC will make our phones the offline shopping credit card of choice. Perhaps. It’s also quite likely that it’s really hard to sort through lots and lots of different items on a tiny screen. That problem won’t be resolved until human eyesight improves and fingers get pointier.

We know that something like one third of all smartphone searches carry local intent, that is: the user is trying to find a physical location. For businesses that are online only these users are the wrong group on which to spend advertising dollars.

To advertisers with a wide brick and mortar footprint the online conversion metrics may not tell the whole story. Perhaps these folks are converting offline instead and the value of the traffic is materially understated?

It is worth testing with POS coupons to get a sense of this, but in the meantime, you might at the least restrict the degree to which that online to offline assumption is being spent within a reasonable geographic proximity to your bricks. Spending beyond online ROI metrics on smartphone traffic 100 miles from the nearest physical location is likely unwise.

Context Is Coming

If marketing hasn’t become sufficiently complex yet, might we think differently about our bids and even ad creative to smartphone users whose device is motionless vs moving 2 mph vs moving 30 mph vs moving 60 mph?

We must crawl before we walk and run before we fly, but SEM isn’t getting any easier any time soon.

None of these refinements will make much difference unless and until the mobile experience we offer our users improves. The average quality of mobile user experiences is low, and it is likely that there are heavy penalties coming soon for brands with poorly optimized mobile user experiences. There are rumors floating around that the penalty for high bounce-rates for organic listings in mobile and mobile QS in ads is about to be increased.

Conclusion

The big picture of all this is simply that we must not expect all users to be equally valuable to our business regardless of the device they’re using and the context in which they’re using it.

We need to market to them smarter, now, and build better experiences for all users in the near future. The growth rate of searches from devices other than desktops and laptops suggests that the penalties for ignoring these differences will only get bigger as the year progresses.

The earliest adopters of sophisticated paid search practices were those companies with deep roots in direct response marketing.

Roots in Direct Response Marketing

In e-commerce, it was the traditional catalog companies that figured out search first. The cost of a book (catalogs are called “books” by insiders) in the mail looked a lot like the cost of a click, and conversion rates looked a great deal like catalog response rates.

Catalogers were steeped in notions of segmenting data by source, by geography, by season, by interaction history, as well as creative and offer testing. They were also comfortable driving by that information.

In finance, it was the credit card companies who were quickest to understand the connection to direct mail and applied the same intellectual rigor that served them in those pursuits. Internet pure plays that grew up in tracked marketing were quick adopters as well.

Many well-established companies that relied historically on TV, newspaper circulars, radio and other media buys that are inherently less trackable were generally slower to adapt to advanced SEM practices, because driving marketing decisions by the numbers wasn’t deeply ingrained in their culture.

(Note: I’m casting no aspersions on the brands below as I don’t know how they’ve driven their programs, just giving samples of the types of companies I’m talking about.)

Roots in Media Buying Not DR

Big Box retailers, banks, clothiers, real estate brokerages and others often still make the mistake of thinking of paid search the same way they think about print ads: impressions, audiences, context and creative.

We often find horrendously broken programs, spending money with no rhyme or reason as to how the money is spent. Mismanagement of the critical non-brand spend is often hidden by brand advertising performance.

In many, many of these cases the objective may be brand building. In other cases like those below, building brand awareness is the only realistic objective.

Purely Brand Building

Where sales are widely distributed and untrackable (Budweiser, Levi Strauss, Ford, Blistex, Paramount, the Florida tourist bureau), or not the objective at all (a presidential campaign, Exxon, BASF), branding is the right way to advertise.

But if brand advertising is the objective, there is no reason the same paid search rigor won’t prove useful. Simply spending a budget, targeted at a blanket Cost per thousand impression (“CPM”) or targeted to capture a certain “impression-share” is not the way to get the most bang for the buck.

How Should Value Be Determined For A Branded Campaign?

For a brand building program, careful consideration should be given to how value is determined. I’ve heard people compare the CPMs available in search or display to print or TV and thought: huh? Talk about comparing apples to oranges!!!

The cost of the impressions may be similar, but what about the value of the impression to the brand? Certainly watching a 30 second TV spot has a different impact on a person than does a text ad that appears on a webpage along with 20 other text blocks.

The human user searching for something relevant to your brand is far more valuable than the regionally/demographically targeted TV viewer (who may be getting a snack during your spot), but the impression itself conveys much less about your brand.

Paid search is managed well when the cost of advertising is married to the value of the advertising. We must first consider carefully the value associated with different types of brand interactions.

A paid search ad may have a very well understood and easily measured cost, let’s say $50 for a day, but what you get for that cost may be a complex set of brand interactions that look like this:

How do we make sense of that value?

The best approach is to create some sort of common value metric hopefully based on data, but minimally based on good intuitive reasoning. It might look like this:

Now, this is a bit idealized, but you get the notion. In this case, the value appears to be worth the costs and that ad can bear a CPC of $0.63. Text ads with similar numbers of impressions may have dramatically more or less value based on the quality of interaction after the exposure. Evaluating by CPM is the wrong way to think about this.

The complexities don’t stop there. For State Farm, Century 21, or Talbots, shouldn’t the brand advertising be more valuable in close proximity to the brick and mortar presence, where dropping by the store or office may become more top of mind and more likely?

Shouldn’t we place even more branding value on traffic coming from mobile devices near the physical locations? Maybe smart phones > tablets > desktop?

If there is more valuable brand advertising available than you’ve budgeted for, how should we prioritize? We may find that we can spend the whole branding budget on very tightly aligned keywords.

Is it better in this case to spend the whole budget on the highly targeted tail, or on a handful of head terms that are less well targeted to your brand?

If you’re the Romney campaign, for example – which type of keyword would you spend the most money on: “Obamacare”, “Health Insurance Reform Act”, or “Health care policy debate”?

During the primary, you may want to appeal to the base more than others and focus on people searching for Obamacare. During the presidential campaign, maybe you don’t need to advertise as much to them, but really want to hammer the “Health care policy debate” searchers on Google, Bing, and YouTube.

The point is simply this: driving brand advertising dollars ‘by the numbers’ pays dividends just as it does with direct response. The complexities of brand value measurement make applying this rigor more challenging, but every bit as worthwhile.

It is not at all uncommon for senior management to become unhappy with a terrific and progressing paid search program because of the way we communicate results.

Averages and aggregates lie, as I’ve argued many times, yet they are also an essential tool of managing a paid search program. Even the manager in the weeds cannot look at the most granular performance data and make any sense of it.

We have to look at data aggregated over time, by category, by subcategory, by match type, by geography, manufacturer, the list goes on and on. To really understand what’s going on, we have to study the data six ways to Sunday to get the complete picture, but that can create problems for reporting out results to senior management.

Management doesn’t have time to study the details and aggregated reports can make great performance look mediocre or worse.

Four Reasons Management Can Be Unhappy With A Great Program

1.  Judging the program by metrics outside of the paid search manager’s control.

The most prevalent example comes from blending brand/navigational search in with competitive non-brand search. Since brand search is almost entirely driven by other marketing efforts, the search manager should neither receive credit nor blame for those numbers. I detailed this one a couple months ago, so I’ll leave it at that.

2.  Management sets goals that aren’t grounded in reality.

Any good SEM firm can point to plenty of examples of putting up gaudy growth numbers when taking over poorly managed programs. But there are also instances when new senior managers come in with “ambitious” goals for a program that is already hitting on all cylinders.

“We’re planning on 100% YOY growth this year, so you’re going to have to work really hard!” Failure to understand market constraints can make excellent performance look “way below our targets.”

3.  Looking at the wrong numbers on the report.

Less common now than in the past, but still not uncommon: people who don’t get search will often latch onto metrics that are indicative of nothing. For example:

• What’s happened to our CTR?!?
• Why is the conversion rate down?
• Our average CPC has gone up! That’s terrible! What’s going on!
• Our average position dropped, someone’s head must roll!
• How many keywords have you added this week? Get cracking!

All of these can be useful diagnostic markers for identifying potential problems and opportunities, but many folks still mistake these for KPIs. They aren’t, and certainly aren’t for management level. A paid search manager might add a bunch of ads that perform differently in the marketplace than others and still be a tremendously helpful addition.

Consider the following example, where the non-brand search results typically look like this in aggregate:

The intrepid search analyst finds a treasure trove of great new keywords and launches them, producing the following results:

Uninformed upper management goes nuts:

“Haaaaaaaaaaarumph! Click-through rate is down 50%! Conversion rates have dropped 4%! CPCs are up nearly 50%! Heads must roll!”

Focusing on the more meaningful available metrics would yield a different response. “Wow! Sales are up over 130% and the ROI actually improved! That’s fantastic!”

4.  Effectively hiding the most meaningful numbers from management.

This manifests itself in many different ways. Let’s consider the case of a major auto insurance company. They take leads in exchange for rate quotes. The average lead is worth $250 to them.

The outstanding search marketing manager conducts research on the back end performance of leads and learns that the leads from certain types of keywords (or geographies or whatever) are significantly more valuable than others (say $500), and that leads generated from searches for “cheap/discount/low-budget insurance” are worth considerably less (say $150).

Looking at the past data this way reveals that they’ve been overspending in some areas and underspending in others. So, she adjusts the targets to generate better ROI for the company.

But senior management wants her fired because from their vantage point, the numbers went south big time.

“Spend is up, leads are down, and the CPL is way out of line with our target!”

Understanding the actual lead values, it’s clear that in fact the search marketer has made the right moves for the company, but if management doesn’t see those numbers…

This phenomena can take many different shapes and forms. In ecommerce, different product lines have different margin structures making a $100 purchase in a high margin category much more valuable to the firm than a $100 purchase in a low margin category.

If the actual margin data can’t be fed into the system dynamically setting different efficiency targets by categories is the next best solution, but that leads to difficulty interpreting the numbers when aggregated across categories.

In this example, the first week’s data reveals fine results, in line with efficiency metrics necessary to maximize operating margin from each category.

The next week sees a seasonal spike in interest in the high margin products and a decline in interest in the low margin items. The well managed program hits the category efficiency metrics again, maximizing operating margin for the company, but the aggregate results showing only costs and sales seems to indicate a decline.

It’s important to note here that from the aggregate numbers there is no way to tell whether the program has gone downhill or continues to perform well. The same aggregate numbers could in fact reflect poor management and missed opportunities.

This is not the fault of upper management’s misunderstanding as much as the difficulty inherent in communicating the complex truth, which is particularly challenging when we can’t present the numbers to which we’re actually driving (net margin in this case). Many companies can feed margin data into a good bid management platform either at check out or in a back feed, but some companies simply can’t tie that thread.

As programs become more sophisticated — incorporating lifetime value considerations, what keywords generate more new customers than old, more business customers than consumers, etc. — the communication and evaluation metrics become that much more challenging to convey

Working Towards A Better Solution

There is no perfect solution here, but there are a few approaches that can help:

• Getting everyone in senior management to have a basic understanding of what the channel can and cannot do, what controls we have available, what the goals of the program are, and the challenges of viewing aggregated data is a great start. The more those folks understand about how search works the more they’ll appreciate the good work you do. Keeping management ignorant can help the incompetent manager, but creates jeopardy for the supremely competent.
• Sharing more detailed reports to management that better reflect all the nuances. This is a great solution if the management team will go along with and actually review more complex reports. The challenge is many managers operate under the dashboard principle: “I just want to see either green numbers or red numbers. Don’t tell me it’s more complicated than that!”
• Determining and reporting out an “all-in value” metric that is consistent with the program’s management goals. If the goal of the program is to drive gross margin dollars, then the aggregated reports should show gross margin dollars, not just sales. If the actual dollars can’t be determined establishing a best proxy can make sense. If experience and research suggests that the advertiser gets different types of leads which have inherently different value,

For example: homeowner’s insurance vs car insurance vs personal property insurance vs motorcycle insurance, and that within those leads of differing value there are classes of keywords within each product that drive leads of different value (homeowners insurance from Beverly Hills vs Tumbleweed, Texas), then assigning and reporting out some approximate dollar value tied to each lead collected will give a much truer picture of the health of the program week-to-week than will aggregate lead counts that treat them all equally.

This notion can be extended, to include a dollar value associated with new customers vs existing, different customer types, non-conversion success metrics like email sign-ups, etc. Assigning assumed $ values does carry the potential pitfall that the folks in accounting look at the report and flip-out because the dollars reported aren’t “real” in their minds.

We shouldn’t avoid complexities because of the communication challenges they present. Raising the bar demands a nuanced approach, and it is incumbent on both the paid search managers and senior management to make sure the lines of communication and the reward structures are aligned to encourage raising the bar, not lowering it.

Werner von Heisenberg courtesy of Wikipedia

In 1927, Werner von Heisenberg documented what he referred to as an “Uncertainty Principle” governing quantum mechanics.

The Uncertainty Principle holds that an observation cannot precisely reveal both the position of a particle at a point in time and its momentum. The more the observation reveals about one, the less the observer can know about the other. A similar principle governs paid search.

Physics sidebar: feel free to skip! If memory serves, the notion is that the act of observation impacts the object. To observe anything — by sound echo-location, sight, touch — we have to bounce something off of the object we observe.

For big things, this is irrelevant. Shining a flashlight on a tree doesn’t impact the tree. However, in the world of subatomic particles, bouncing photons or anything else off a tiny particle has a big impact.

If you want to know where a particle is at an instant in time, you have to “hit” it hard to get the answer quickly, but in doing so you impart a huge and unpredictable change in momentum to what you’re trying to observe.

The Paid Search Uncertainty Principle

This reads: “The amount of variance tolerated in Advertising Efficiency (“E”) times the amount of variance tolerated in the volume of advertising spend (“V”) is greater than some constant, K.” As the variance in one approaches zero, the variance in the other approaches infinity.

Okay, okay, enough of the physics metaphor, the idea is this: you can’t control both spend levels and efficiency metrics. The more you predetermine one, the less control you have over the other.

This fundamental law of paid search may be its least understood, as many companies determine an efficiency target and then simultaneously fix a rigid budget for how much they will spend in media. Those companies are often frustrated with their results.

Fixing a budget for ad spend translates to the following:

“I’m going to spend this amount of money regardless of market opportunities. I will flush money down the toilet if I must in order to spend my entire budget. If, on the flip side, the market conditions are such that I can make $20 in profit for every $10 I invest, I will nevertheless stop spending money when I reach my budget.”

Fixing an efficiency target translates to the following:

“I am going to spend my advertising dollars at an efficiency that makes sense for my business. I will spend an unlimited amount of money if I can do so efficiently. I will spend no money if market conditions dictate that I can’t spend efficiently.”

Stated as such, the Uncertainty Principle clearly follows.

Yet enterprise-level advertisers try to violate this principle all the time:

“We’re going to spend $1 Million this month, and we need to see an ROI of 5 to 1 (or a CPL of $30).”

It’s nice to have goals, and it’s smart to try to forecast what you’re likely to see, but at the end of the day market conditions may dictate that you pick one or the other target to hit, or you’ll likely end up missing both.

Important caveat: this assumes that your program is well managed. We’ve taken over many horribly mismanaged programs and both grown the spend and greatly improved the ROI.

But all things being equal, for any well managed program there is a trade off between efficiency and volume that has to be understood.

Paid search managers have control over many pieces of the game:

• The user searches that will and will not fire an ad
• The text of the ad fired by a particular search
• Where ads are served (both domains to a greater or lesser degree, and geographies)
• On what devices they are served
• The maximum amount the advertiser pays for a click on a given ad under almost any circumstance

As I’ve argued relentlessly in the past, we’d like more and better controls over many of the above, but we certainly do have a large degree of control.

But it is also important to recognize those factors paid search managers do not control:

• The volume of user search in a category. If the users don’t search, we can’t serve ads.
• How much other advertisers are willing to pay for clicks. If other advertisers behave irrationally, or change their bidding practices significantly, it will have a material impact on the marketplace opportunities.
• The advertiser’s selection, pricing, content quality, and promotions are often outside of the control of the paid search manager. These impact the value that can be extracted from traffic and thus the amount the paid search manager can spend for that traffic.
• Competitors’ promotions, selection, content quality, pricing, etc. Significant changes in the business practices of competitors impacts CTR, QS, and conversion rates for their ads and thereby shifts the Volume v Efficiency landscape in ways we can’t control.

The point of this isn’t to make excuses. Indeed, if the goal of enterprise search advertising  is to spend a large budget, there is no excuse for spending significantly more or less than the budget. If the goal is to hit an efficiency metric (CPL, ROI, Margin to cost ratio, whatever) for non-brand search, there is no excuse for missing that target.

However, introducing two targets often creates an unsolvable problem for paid search managers of any size.

Hitting both an efficiency metric and a volume goal requires good execution, yes, but it also requires favorable market conditions over which the paid search manager has no control.

Paid search managers must be held accountable for those elements that are within their domain, but holding folks accountable for conditions outside of their control is both unfair and unwise.