• The more specific keyword had a higher bid, hence we ended up paying more for the traffic than it was worth to us.
• The landing page was less targeted, so we were taking bad traffic and landing it on the wrong page, making it even less valuable traffic.
• Because of the poor quality traffic pouring in on what had been a high quality term, we bid that term down, meaning we also get less of the high quality traffic that the term normally draws.

Our reps at Google at the time told us that this couldn’t happen, that the exact matched keyword would always get precedence over the broad mis-match so what we were seeing… er… “wasn’t happening.”

We knew we were right about the phenomena and given their protestations that exact matches always won we suspected it was a mistake on Google’s part. We asked them if there was logic that makes exceptions to the exact match precedence if the ad is paused. They said “yes.” We then suggested that back in the day when there was a minimum bid, that an ad bid below that minimum might also be considered paused. They concurred. Then we suggested that when the minimum bid was replaced by the first page minimum bid perhaps the code wasn’t updated and any ad that didn’t meet that minimum would be treated as “paused.” They said at the time “That shouldn’t be the case; that isn’t what we intended; if it’s a mistake we’ll fix it.”

Googlers in good authority now tell me that that bug doesn’t exist—there’s no reference to the first page minimum in the code that drives the rankings. Instead the explanation is simply that the more specific keyword must have a higher Quality Score than the exact matched more general keyword… or that its combination of bid and Quality Score are higher than the exact matched term. We’re told that this shouldn’t really happen if the Quality Score of the generic ad is good, but our data suggests otherwise. Just a cursory look at our data showed plenty of instances where an exact match ad with a Quality Score of 10 was passed over for a broad matched ad with a higher bid.

Managing this self-competition with the current tool set is not just cumbersome, it’s impossible to do well. Adding every keyword as an exact matched negative for every other keyword in the account is unworkable. Bombing in all the general keywords as exact match negatives for all the more specific keywords is doable but time intensive and therefore costly.

Some advertisers simply give up on broad match to prevent the shenanigans. We’d say that’s throwing out the baby with the bathwater, but understand the frustration. We do what we can, but anyone claiming to have this problem solved is either delusional or deceptive.

As I noted a couple of years back, one way Google could effectively self-destruct would be to go too far down the path of allowing higher bid/lower CTR keywords to take precedence over the right keywords. As described above, not only does spraying the traffic around unpredictably make all bidding systems less efficient and hence spend less money to reach the same efficiency target, the bigger danger is in alienating the shoppers who use the ads.

Advertisers bid ads down for good reason, like when inventory is thin, and if other ads take their place and continue drawing in traffic that won’t convert it is a disservice to the user as well as the advertiser.

Sometimes I rant around the office saying things like: “Advertisers should sue! Here we’ve given Google instructions as to how much we’re willing to pay for people who type in “Foo Bar,” but when someone types in “Foo Bar” Google decides to serve my ad for “Left-handed steel foo bar” which has a much higher bid! That should be illegal!!!”

Important note: I should be ignored when I rant on like this—other times, too, but particularly when I rant. Anyone looking at the AdWords terms of use for ten seconds would realize that Google can serve whatever ad it wants so there’s no point in calling a lawyer.

Google’s engineers genuinely believe they can algorithmically pick better ads to serve than the advertisers can. This may be true for badly managed accounts, but is not true for well-managed campaigns. If this notion that sometimes humans are smarter than the machines is offensive to engineers, perhaps it could be framed in the language of “crowd sourcing.”

If the engineering team is willing to acknowledge that some folks might actually choose ads, landing pages and bids rationally, there may be a profit maximization angle as well. Google is not “evil;” it is a publicly traded company looking to grow its top and bottom line just like us. My argument here isn’t that they can’t do this legally, nor is it that they shouldn’t do it ethically. The argument is that this isn’t a good business decision on their part.

Bing’s path to victory lies not in stealing Google’s organic traffic, but in taking Google’s shopping traffic. That’s what Cashback is about, and if Google places short term revenue maximization over long-term ad relevance they’re opening the door for Bing to step through.

If average users decide that “Google is great for research, but go to Bing for shopping” Microsoft’s big investment might just pay off.

Well, no.

Volume and efficiency are always at tension and how that tension is balanced is highly revealing.

There are two fundamental approaches to this problem:

• Spend the budget and “maximize the ROI.”
• Hit the efficiency target and maximize the budget within that constraint.

We have always believed that the ROI target should take precedence. Let’s say Acme has a budget of $100,000 and an ROI target of 4 to 1. The first approach: Spend the $100K and maximize the ROI.

Scenario 1: Spending $100K leads to significantly lower than 4 to 1 ROI:

Does it make sense to keep spending when the ROI turns south? The law of diminishing marginal returns tells us that next dollar of additional advertising wisely spent will always return less than the previous dollar. When we hit the point that the next dollar returns less than the target efficiency, or less than break even, why would we keep spending? If the returns are poor it’s like feeding money into a shredder simply because the goal was to “get rid of” some fixed amount of money.

Scenario 2: Spending $100K leads to significantly higher than 4 to 1 ROI:

Does it make sense to stop spending money if the ROI is great? When the delay between investment and return is long or uncertain, budgets are essential. For advertisers taking direct sales the paid search returns happen before the media expense is billed, so the cash flow argument doesn’t carry much weight. And, when the returns are cash rather than leads, there isn’t even uncertainty as to the value of those returns.

It’s like having a commissioned sales force but telling them to go home when they’ve hit some salary cap. If the commission rates make sense, why would anyone apply the brakes?

“Maximizing ROI” implies exactly this kind of budget first mentality.

When client prospects approach us with budgets that are inconsistent with their ROI objectives we tell them that up front. Imagine a company selling all varieties of bungee cords and hoping to spend $250K a month at a 5 to 1 ROI. We’d tell them that given the market opportunity we can only do one or the other, and would recommend pursuing the 5 to 1 ROI as far as the market will allow.

Most other agencies would say: “Those are aggressive goals, but we’re up for the challenge. We’ll spend the budget and see what we can do about the ROI.” Next month, the discussion will center on how they can “maximize the ROI” while still spending like a drunken sailor. That’s good for the agency’s revenue, and good for the engines, but I’m not sure it’s in the advertiser’s best interest.

Maximizing ROI isn’t necessarily even the most profitable approach. By definition ROI is a ratio. A 5 to 1 ratio seems better than a 4 to 1, but looks can be deceiving. Let’s say the advertiser has 40 points of margin on a sale. On $10K in advertising the advertiser generates $50K in sales. $50K in sales yields $20K in margin, less the advertising means $10K in marketing income. Cool.

But if the additional aggressiveness of a 4 to 1 target gets them significantly higher on the page and hence more top line, the extra volume may generate more marketing income. Depending on the bidding landscape, perhaps $20K in advertising well-spent generates $80K. That 4 to 1 ratio, generates $32K in margin and $12K in marking income. A smaller ratio, but more money, and as every banker knows, you don’t put percentages in the bank.

According to the square root rule, profit is maximized when the ratio of margin generated to advertising spend is 2 to 1. In practice we find running that lean cuts the top line too much for dollar maximization. Instead, we’ve found that a ratio of margin to advertising of 1.4 – 1.7 tends to maximize marketing income.

The “spend the budget first” mentality does make sense in some contexts. When the return on investment is uncertain (brand building) or delayed (lead generation), budgeting is wise. Too often, though, budgets are used simply because they are customary.

For this channel, any company with measured, immediate ROI goals should place those goals first and let the ad spend rise and fall with the market opportunities.

Language matters in paid search. The language agencies use in describing themselves can reveal a great deal. “We’ll maximize your ROI” screams: “We think about spending your budget first and hitting your efficiency targets last.”

How many marketers think sending an identical email to everyone on a mailing list will perform better than individually tailored emails based on each recipient’s past behavior?

How many think sending non-personalized direct mail pieces will outperform personalized outreaches?

Anybody? Anybody? Bueller?

If no one thinks the monolithic approach would ever beat the segmentation approach, why is it that some PPC professionals advocate ditching long tail keywords and letting broad match catch the tail?

The benefits of the long tail search advertising are exactly like the benefits of email segmentation. Consider:

The more targeted ad is served having better ad copy which raises CTR and therefore Quality Score. This leads to more traffic from the same number of user searches.

A more targeted user search necessitates a more targeted landing page. Not the somewhat generic “Fender Guitar” page, but rather the “Fender 4 String Bass” page. That more targeted page will likely deliver a higher conversion rate, hence more sales from the traffic generated by the ad.

Folks looking for “Fender 4 String Bass” will have a much higher conversion rate than those searching for “Fender guitars.” That fact means you can bid higher on the more targeted ads, putting the ads for the most targeted traffic higher on the page than the more general terms which could be yielding more impressions and higher CTR.

With more of the highly targeted traffic going to the tail keywords the less targeted, more general keywords will be bid down to more accurately reflect the value of the traffic. As I mentioned in my post on syndication partners, better resource allocation leads to a bigger program at the same rate of efficiency.

The advantages of a thoroughly developed keyword list are obvious in principle, but does this really matter in practice?

The answer is an emphatic “YES!”

In a recent empirical study, we took apart campaigns from a number of our clients ranging in size, product types, price points and more, to see the degree to which the “long tail” matters. We found that the importance of those targeted searches varied tremendously from a low of 8% of the total business to a high of an amazing 83% of the total. The median of the study group was 31%.

Obviously if the program is tiny, and the tail is only 30% of the game, it might not be worth the attention, but for large programs even 8% is worth chasing. In the case above that 8% represented over $200K in sales per month.

Nevertheless, some argue that those targeted searches would be captured by broad match anyway, so there’s no real advantage gained by building out and maintaining a comprehensive keyword list.

Not so.

As a test case, we took a client in the consumer electronics space who carries products from more than 1,000 different vendors. We then asked the question: what’s the traffic value differential between the highest traffic keyword for each vendor brand, and the rest of the keywords associated with that brand? In other words, if we just ran the highest traffic term for each brand on broad match to what degree would we be blending together traffic of vastly different quality?

The answer: “We’d be mixing apples and oranges.” In the test case and for vendor brands with enough traffic to have a tolerable signal to noise ratio the median variance in traffic quality (margin dollars per click) between the brand-specific “head” and the brand-specific “tail” was 84%! And not always in the direction expected.

Throwing all the granular data into the head keyword performance will very often result in significant over spending on one cohort of user searches and significant opportunity lost by under spending on the other cohort(s) of traffic.

So, there isn’t much question that for a substantial program the tail matters, and treating the head keywords the same as the tail keywords leads to significant mishandling of the bids.

The other rationale for ignoring the tail relates to the cost of building and maintaining long tail campaigns. “The time is better spent elsewhere.” Well, that could be true if either of the following is true: 1) it’s a small program with an insignificant tail, not worth the management cost to go after, or 2) you lack the power tools and algorithms necessary to build and manage the tail comprehensively with cost-effective human effort.

There is no solution for the first issue. If the whole program isn’t worth spending much time on, then certainly the tail isn’t worth it.

The second issue however is simply a resource and know-how constraint, which for a professional paid search manager should not be there. Telling folks “we don’t have the tools to manage your program effectively” is a hard message to deliver, but it’s the right message if it’s true. Telling clients: “The tail is unimportant,” or “The tail is effectively handled by broad match” simply isn’t honest for most sizable programs.

In competitive paid search, the wrong landing page will cost you money. While choosing the right pages has always struck us as fairly intuitive, we think that’s not the case for most other paid search marketers, given the widespread phenomena of poor landing page choices that we’ve seen. It isn’t that paid search managers have lost their intuition, its that they’ve turned landing page choices over to the same machines that pick their keywords.

Because agreement between the keyword and the landing pages is so crucial, we have always put tremendous time and thought into their development. Ours is a human-centric process that involves sharp people working from the landing pages backwards to pick keywords. We built proprietary tools to speed the process and build out comprehensive lists, but it all starts with our analysts using good judgment, knowledge of language and leveraging years of experience with what makes a proper landing page.

What we’ve learned from thousands of tests for national retailers:

The Platinum Rule: Landing pages should reflect the depth of the user search. The goal is to show the user the widest selection of product that responds to her/his search, and nothing more. Someone searching for “oxford shirts” should not land on the homepage, unless oxford shirts are all the advertiser sells; but just as important: neither should they land on a product page even if it’s a best seller, unless it’s the only one the advertiser carries.

This is the controlling rule, all others are subservient to it.

Product pages are great landing pages for product-specific search terms. Nothing else should land on a product page. The notion that “closer to the shopping cart is better” is just dead wrong. If more than one product responds to the search then the landing page must reflect that, otherwise the user’s impression is, “This is the only one they have.” You might think “they’ll realize we have others and sniff around,” and some will, but too many will hit the back button and look elsewhere for a broader selection.

Search results pages usually beat sub-category pages on more general searches, but only if the site has robust search results. Be particularly careful about sending “everything” to the search results pages, as synonyms and misspellings often aren’t handled well by internal site search boxes.

“Show All” option usually beats “top sellers” in search results for sites that have those options.

Creating a special landing page can make sense if your site “forks” on gender or something like it. Someone searching for “North Face Jackets” may be looking for men’s, women’s or kid’s options, and on many sites there is no page that provides that nice set of options. Search results may be the best choice, but creating these “path fork pages” can be very helpful. Onlineshoes handles this beautifully for “merrell shoes.”

Test templates rather than individual pages. Very rarely does an individual landing page get enough traffic to generate statistically significant test results in a reasonable amount of time. Testing the concept of sub-category vs search across all relevant pages provides actionable information that credibly applies to all the individual pages. Intuition is not a substitute for testing, but it certainly provides a time-saving short-cut to the right answer!

Robots pick lousy landing pages. We sometimes supplement our own tools with other Keyword research tools, but do so cautiously. Our analysts find that they need to change the landing pages from those suggested ~70% of the time. Those who take the easy route and just post what the robot spits out are losing money.

Let’s look at this last point more closely.

Using Google’s Search-Based Keyword tool I looked for keywords for Landsend.com (not one of our clients… yet :-)

The keywords themselves were predictably dangerous: “7 jeans,” “thinsulate,” “shearling,” but I’ve already blogged about that, and the new tool does do a somewhat better job of screening out the junk.

The problem with respect to landing pages is that the tool picks product pages for general search terms. For the phrase “moccasin shoes” Google chooses this page, when this one will generate a much higher conversion rate. That sub-category page could be tested against a search results page, and the results might be interesting, but the SKU page is certainly wrong.

Here are the first 20 keywords suggested for Land’s End with the type of landing page and my assessment of how likely that page is to be the best choice.

As I see it, just six of the 20 keywords were the both the right keyword and the right landing page. Not so good.

To be clear, I don’t fault the sharp folks at Google. Trying to determine the optimal landing page often requires judgment, and the judgment here may defy algorithmic solution. As I mentioned, search results pages often produce the highest conversion rates, but there is no way for Google to determine what an advertiser’s search URL structure looks like nor whether those search results are both well-presented and fairly robust.

Keyword and landing page tools may save time versus doing things manually, but used indiscriminately—as they often are—they hurt conversion necessitating lower bids and a smaller program.

Many have found it hard to raise conversion rates for retail websites through site design changes. Multivariate testing is often a long series of null results. However, choosing the right landing page is a different issue.

The landing page responsive to a search result matters less in its design than in its content. If the user isn’t exposed to the right content—however it may be arranged—they will hit the back button and the paid search program will suffer.

Fact: The quality of traffic coming from a search engine itself is of significantly higher quality than the traffic coming from its syndicate networks.

This has two important consequences:

1. Bidding the same for traffic on the Google syndication networks as on Google.com leaves opportunity on the table.
2. By not allowing advertisers to differentiate bids based on the quality of traffic driven by each network partner, Yahoo and Google lose money as do the advertisers.

Theory

Let’s take some basic paid search principles and show how they apply to the syndicate partners. I’m going to make some simplifying assumptions to prevent us from getting mired in detail and missing the bigger picture.

The value of traffic in paid search varies terrifically by the user’s search and the ad it fires. Since we can’t really bid based on user search, let’s see what that looks like as a function of keyword:

Cool! Now let’s say the advertiser is willing to put 80% of the true value of the paid search income back into the marketing to drive the top line but still make some immediate profit.

Ideally, the bids would look like this:

By matching the bids to the value of the traffic on each keyword we generate as much traffic as we can afford from each keyword, thereby maximizing revenue within our efficiency requirements.

There is, however, another way to achieve the efficiency objective. Namely, by bidding to the average value of the collection of keywords.

In this simplest of models, assuming that the underbidding and overbidding don’t move the average value of the traffic—a big assumption and probably never the case—you still end up spending 80% of the paid search income on marketing, which was the target all along.

So if either way you hit the desired efficiency target, what’s the difference?

Volume. By wasting money on some terms, and missing opportunities for more sales on others, the poor bidding methodology hits its efficiency objectives but does so at the expense of sales volume.

Important corollary: Lower sales volume at the same efficiency by definition means lower advertising revenues for the engines.

Applying theory to the syndicate networks

So, if what we’ve learned is that applying the same bids to traffic of varying quality leads to fewer sales/leads for the advertiser and less revenue for the engines how does that effect our view of the syndicate partners?

Let’s look at some data. I grabbed data from a handful of our clients and took median values by referring domain for click volume, and order volume as a percentage of the total traffic for each client. I also calculated median values for how each domain’s conversion rate compared to the average conversion rate for that ad network. The comparative conversion rate data is most interesting.

Note: sites are ranked from left to right in descending order of traffic volume for the median client.

Notice that traffic from Google.com, AOL, Amazon, and Comcast tend to send much higher than average quality traffic: 16 – 20% better than average. At the same time, eBay, Ask and the comparison shopping engines, tend to send significantly lower than average quality traffic.

For AdWords advertisers can and should bid differently on the Google.com domain and the rest of the network. Looking at these medians as if they were one advertiser’s data, we’d end up bidding the Google.com-only version of the account up by 16% and the Google.com + syndication network version of the account down by 33% (Google.com traffic is slightly more than 2/3 of the total, hence the disparity).

My sources tell me we’re among the very few folks in the space doing this, which is surprising as we’ve been doing it and advocating it to others for more than two and a half years.

This two-tier approach will help materially, but is in no way ideal bidding. In this, we end up underbidding on AOL and Amazon traffic, and still overbid on eBay, Ask and others. Better, but not great.

If we take a look at the same type of data from the Yahoo network we see even greater disparities.

Yahoo.com traffic is 35% better than the average Yahoo ad network traffic! The rest of the network averages about 42% below the aggregate average! But, even among those there are winners and losers.

Yahoo does not give us the opportunity to bid differently on the network, but does give us the opportunity to exclude traffic from particular domains. Given the lower traffic volume on Yahoo, it can be difficult to separate signal from noise to identify those domains that send particularly low quality traffic. Sometimes looking across multiple accounts helps us spot trends that we can’t really see looking at a single account’s data.

But excluding sites isn’t really what we’re after either. In most cases the traffic isn’t worth nothing, it’s just worth less. Paying the right amount for it would allow us to generate sales cost effectively from each domain in the syndicate network and most importantly, would allow us to push the gas harder on the high quality traffic provided by some.

Both Google and Yahoo claim to discount the CPCs from network partners, but our experience suggests the discounts aren’t adequate. Moreover, since any good bidding system bases bids on expected revenues, rather than sunk costs, the discounted CPCs wouldn’t solve the problem even if they were right. The bids might be right for the bad performers because of the discounts, but we’d still under bid on the higher quality traffic.

The obvious best solution for the advertisers and the engines is to allow us to bid differently by domains. Creating separate campaigns for each domain might be too much to manage, but account level percentage adjustments based on each advertisers data would be easy.

Perhaps the hangup lies with publishers. If eBay’s revenue as a syndicate partner dropped by 40% they might use that space for display ads rather than search, and perhaps they’d rent that space through a non-Google exchange. Legacy revenue sharing agreements between the engines and the network partners may also be a barrier.

Whatever the case, we would like to have more control over what we pay for the traffic we get from the syndicate partners. We hope you folks will join us in calling on the engines to make this change!

Let’s take a look at some of the common trends we see in Q4 data and discuss how to, and how not to address these effects through anticipatory bid management.

We’ll use actual data from one of our client’s first holiday season with RKG to highlight some of the common phenomena. The first holiday season is always challenging, because we don’t have data to use to help us anticipate the timing. As such, this data reveals some missed opportunities that we can model the next time around.

On each chart 100% means the average week between mid-September and mid-October. Lifts above or below reflect the week’s changes from that norm. We’re using “sales dollars” as a proxy for value because it is the most widely used. A better metric is margin dollars with frauds and cancels knocked out.

Newsflash: Sales go up at holiday!

This happens for two reasons:

• Traffic volume increases. More people search, more people buy => higher PPC sales and advertising costs.
• Shopping behavior changes. Not only are there more consumers in the market, their propensity to buy and how much they buy fluctuates throughout the period.

It is this second factor that allows us to bid differently, and makes anticipatory bidding imperative. An increase in the volume of traffic by itself does not change the value of the traffic or what we can afford to pay for each click; it is the changes in the traffic’s value that allow for bidding adjustments.

So, yes traffic volumes spike:

But it’s the change in conversion rates and average order sizes that is interesting and actionable:

(There is undoubtedly social commentary to be made over the fact that AOV dips at gift giving season—we spend more on ourselves than on our loved ones?—but it’s also important to recognize that discounting leads to thinner margins which may in turn demand different efficiency targets.)

When we roll these phenomena together we get the true picture of changes to the traffic value over time.

If the goal is to maximize sales during the period at the same efficiency as normal then we should push/pull bids proportionally to the changes in traffic value. That way we maintain constant efficiency, and generate the most sales possible. (Folks who budget search instead look to spend the budget at the lowest cost to sales ratio possible. This is a totally different problem, and, we think, the wrong approach to paid search.)

So, how did we do?

Not bad, given that we didn’t have good data to use to help us anticipate the changes. From this perspective it looks like we may have overspent just a touch the first two weeks in November, and may have underspent a bit the week of the Christmas holiday and the week after.

But hold on! This view ties the clicks and costs on a given day to the sales that happen that day. We know because of order latency that many of the orders placed today came from clicks on ads long before. This suggests we might have actually underspent during the ramp up and left opportunity on the table. So, let’s instead tie the orders to the time of the clicks that drove them and see what that reveals.

Well, it’s not an entirely different picture, but it is slightly different. By this view we were pretty much on target for those first two weeks in November, and it was the two weeks after that where we may have left a bit of opportunity on the table. Good to know! We’ll use these insights this year to do our jobs that much better.

Some folks out there will suggest that you should push the gas much harder early on to catch consumers in the “early phase of the consideration cycle.” I’ve heard folks say they’ll increase their cost to sales threshold by 50% or even 100% prior to the real increase in conversion rates, arguing that those “inefficient” weeks will appear very efficient when viewed by the orders those clicks seeded. I say: if the data from previous year supports that, go for it! But we haven’t seen shifts that dramatic, or anywhere close.

One other pitfall to avoid: Black Box bidding systems that don’t allow for anticipatory bidding. We see data that looks like what’s below every year from agencies that allow the algorithms to do it all, and every year the algorithm reacts too late during the ramp up and overspends greatly after the holiday ends.

These folks didn’t fish enough when the fish were biting, and fished too much when they weren’t. Overall, they may hit their efficiency targets, but they won’t end up with as many fish as they would have had with a smart analyst at the controls.

No two retailers will show the same trends, so let your data be your guide.

Here’s to a profitable Q4!

Paid search is an amazingly powerful channel for capturing existing demand. It is not a demand generating channel. Like the yellow pages people only see the ads when they’re in the market for the advertiser’s products or services. Yellow page ads don’t make people interested in taking yoga; they convince people who are interested in yoga to try the advertiser’s studio. Paid search plays the same role, but it’s easy for marketers steeped in traditional channels to miss the implications of this fact.

For demand generating channels planning is of paramount importance. Television ads seek to 1) create awareness of a brand; 2) infuse that brand awareness with positive connotations; and 3) catapult that positive awareness into action. Marketers must plan the proper messaging and imagery for each phase, and plan the timing and budgets. Even direct marketers like catalogers must plan contact strategies, page counts and formats, circulation counts and frequency of touches. Planning is required because the marketer must generate interest and action, and the time lag between marketing investment and eventual payoff is both long and not guaranteed.

Paid search is different in every respect. Budgets don’t make sense for advertisers with shopping carts on their sites because the sales happen before they pay for the ads. We don’t have to plan what products to push or when; the ads will be served to the right audience at the right time and in the right volume to respond to consumer demand.

This is not to suggest that paid search requires no work. Paid search demands well honed processes for dealing with product turnover, landing page changes, promotional offers, seasonal shifts and more. Skilled analysts study data six-ways-to-Sunday to find nuggets of gold in search logs, in click-to-order intervals, and in new classification schemes that feed statistically meaningful differentiators to our algorithms. Reacting to shifts in consumer behavior and anticipating them whenever possible takes knowledge, skill and mountains of well-organized data.

However, it doesn’t really require a plan. Indeed, predicting what projects will be most valuable 30 days from now, or 60 is both impossible and unwise. The priorities change depending on all kinds of factors outside of the analyst’s control. Locking an analyst into an arbitrary schedule prevents them from adjusting to the priorities of the moment and thereby squancders opportunities.

Asking an analyst for a long range plan for a paid search campaign is like asking your stock broker what stocks you should buy and which you should sell 6 months from now. They can make those recommendations if you really want them, but usually you’d be wise not follow them.

Absence of “a plan” is unsettling, particularly when you’re working with an agency. You can’t see them hard at work at their desks every day, so it’s natural to wonder whether they’re asleep at the switch. This fear is compounded by the reality of PPC agencies that all too often are asleep at the switch.

Trust comes from results, and in PPC results come from smart people with great tools and great processes. Anticipating and reacting to changes in the landscape requires hard work, knowledge, skills and the flexibility to follow one’s nose.

Problem: Successful tests can look bad initially because of order latency

For example, let’s say the order latency for a particular advertiser with a 14 day cookie window looks like this:

56% of the orders come in within 24 hours of the time of the click, 10% come during the next 24 hour period, etc.

So, on that first day you see 100% of the clicks on your ads, but not nearly all of the orders those clicks will drive—actually, quite a bit less than 56%, as clicks late in the day have less time to “mature.” For sake of simplicity, let’s ignore that bit. Doing so allows us to map out what a tremendously successful test might look like.

Let’s say an advertiser launches a new product category and new keyword ads are developed. Let’s say the advertiser’s efficiency target is a 25% cost to sales ratio, and let’s say their brilliant PPC firm nails the bids right out of the gate. The clicks generated on day 1 cost $1,000 and will eventually drive $4,000 in sales, but here’s what the results look like as they unfold spending $1,000 every day at perfect efficiency:

Yielding a day-to-day apparent cost to sales (A/S) ratio that looks like this:

The first few days of the test it appears that the efficiency is way above the advertiser’s comfort threshold. It takes the full duration of the cookie window, for the observed efficiency to match the actual efficiency of the advertising. Advertisers who don’t recognize this effect may cancel tests, or pull back on the rudder too quickly.

Indeed, what this suggests is that every launch and every extra bid push will appear to be less helpful to the top line and more harmful to the bottom line than reality. On the flip side, every pull back on bids will appear to be more helpful to the bottom line and less harmful to the top line than it really is because of the lagging orders from the higher click volumes that preceded the test.

The greater the order latency, the bigger the impact. We typically find that more considered purchases, and higher average order value advertisers have greater latency than average which impacts the proper length for the cookie window.

However, no one wants to wait 14 or 30 or 90 days to read the results of a test. In the example above, the PPC agency hit the bids right on the head from day 1. When that doesn’t happen, it’s good to find out sooner rather than later that you’re undershooting or overshooting.

Two methods for evaluating tests

Shorten the sales window. Instead of evaluating the test based on the full cookie window, study the data based on a same session or one-hour sales interval. In the example above, if 35% of the eventual orders normally come within the first hour of the click, extrapolate the results from the first few days based on that number. If the ratio of cost to (observed 1-hr sales/0.35) is on target, the test is probably on target.

If an advertiser is attempting to learn the top-line vs bottom-line trade off of bidding to a 30% A/S target rather than a 25% A/S target, compare the % increase in 1-hour sales to the % increase in cost. That should be a pretty good proxy for the A/S ratio on the incremental sales.

Tie orders to the time of the click. Most reports show the PPC costs for the day, and the PPC sales taken that day. It’s entirely likely that half of the sales taken that day came from earlier clicks. By running reports tying the sales to the time of the click, rather than the time of the order, you get a much clearer picture of what your actions on that day did for you over the long haul. This is particularly useful for studying past tests and anticipatory bidding at the holidays to see whether you anticipated the improvement in traffic quality appropriately.

The problem with the first method is that it assumes the latency for the new product category, or incremental traffic, will be the same as it’s been in the past. Not a bad guess, but potentially misleading. The problem with the second method is that you can’t use it fully until the cookie windows have elapsed.

By using method 1 during the early phases of the test and method 2 after the test is “complete,” a good analyst can avoid missing opportunities and overspending during the test, and get a dead-eye accurate read on the results after the fact.

RKG is in the midst of a research collaboration with Digital Element and Kevin Hillstrom of MineThatData to determine if some well-known truths from the catalog industry also apply to the world of paid search, namely that geography matters.

Catalogers have known for 50 years or more that people in rural areas respond to offers at a significantly higher rate than those in urban areas. Indeed, postal zones C & D, corresponding to semi-rural and rural areas, have always outperformed zones A & B. Is the same true in Paid Search?

The early answer appears to be: “Absolutely!” Just looking at low population density states like Wyoming, Montana, Alaska, etc, the quality of the traffic appears to be more than 60% higher than that of more urban states. We’re going to take a look deeper along the lines of postal codes to see if this trend is as clear in PPC as it is in catalog mailings.

Another factor catalog mailers have always known: the presence of retail stores matters. Not surprisingly, if you send a catalog full of terrific products to someone who lives near a physical store selling similar products, you’ll drive a lot of sales to that store. If that store is part of your retail chain, great, if not…

Our study will take a look at the impact of having a retail chain store in the same zip code as the searcher. Indeed, this might allow us some insight into the elusive store spillover effect. By comparing the quality of traffic in similar zip codes with and without a physical store presence, we might conjecture that the difference is a pretty good proxy for the amount of spillover.

Kevin Hillstrom has done pioneering work in the field for catalogers. We hope to find out whether the same notions hold true for retail chains and online pure-plays that don’t mail books.

What’s the point? Measuring the phenomena doesn’t necessarily mean we can act on it. Who wants to set up complete campaigns for each zip code?!? No one, and indeed, slicing that thin would leave you with no data to model.

However, we hope that armed with data, we can convince the engines to give us two additional tools—er, beyond the one’s I already asked for—that would allow us to manage programs at the next level.

1. Population density settings. Maybe just 4 levels, corresponding to the postal zones. This would allow us to create at most 5 variants that would capture the benefits, and we might not need that many.
2. Zip Code list tagging. Let us set up a list of zip codes representing anything (our client’s stores, their competitor’s stores, whatever). That tagged group (”my stores”) could be applied to campaigns to either establish different efficiency targets—if I know 20% of the sales happen in my brick and mortar store rather than online I can target a different efficiency threshold for that campaign—or simply suppress ad service to avoid driving traffic to a competitor’s store.

Sophisticated marketing techniques allow retailers to generate more sales for their marketing dollars, and the more sophisticated the tools the more retailers can spend cost effectively. That’s good for the retailer, the engines, and the agencies that handle complex accounts well.

Previously, we tried to put the “PPC Buying Cycle“—touches on multiple PPC ads—into its proper perspective. Turns out it’s a pretty small effect. Unfortunately, many agencies continue to hype the effect as they seem solely interested in having their clients spend more rather than spend wisely. Is the same true with the “Cross-Channel Buying Cycle?” Yes and no.

As we look at the data across a number of multi-channel retailers we’ve found that marked differences in the way consumers use each channel mean that cross-channel interactions have profound impact on some channels and not much on others.

Likelihood of multiple touches

Channels that have the greatest likelihood of multiple touches have the most potential to be impacted by changing allocation from last touch to something more advanced. Our data suggests that the channels most likely to involve multiple touches are affiliates, comparison shopping engines and email.

Consumers who buy after clicking a competitive (non brand) paid search ad are the least likely to have been to the site previously through a different channel. In our research, only 10 to 20% of buyers who touched a PPC ad last came through any other channel previously. Compare this to affiliate traffic, where 60 – 75% of buyers came through another channel first.

This means shifting from last touch to shared credit to first touch allocation only impacts 10 to 20% of PPC orders, while the same shift has a much larger impact on the perceived value of affiliates, comparison shopping engines and email.

Initiators versus followers

If channels were all equally likely to be first as last in multichannel interactions we might find that the net effect of changing allocation schemes is zero. That turns out not to be the case. Some channels are far more likely than others to be the first touch when more than one channel is used.

Competitive PPC is much more likely to be the first touch when there are multiple touches involved. This means that moving credit from last touch towards earlier touches does tend to “help” PPC. Natural search benefits from this same phenomena.

In contrast affiliates are almost always the last touch in multi-touch interactions, meaning shifts away from last touch credit have a decidedly negative impact on the perceived value of affiliate programs. Comparison shopping engines and email tend to suffer as well.

The net effect

What we’ve found is that these two factors together mean that yes, in fact, the perception of PPC benefits from crediting earlier touches in the cycle. However, because fewer PPC orders are in play than other channels—that first effect—the change is smaller than many folks seem to think. Indeed, in our research moving credit from 100% to the last touch to 100% to the first touch, competitive PPC only picks up 5 to 10%. Less dramatic allocation shifts take those numbers down even further.

Shop.org is organizing a group to define standards for credit allocation. I’m going to throw my name in the hat to join said group, but I fear that some of the folks in the group may be more interested in trumpeting the effect than measuring it. We shall see.

In the meantime, I’ll post more of our findings over on RKGBlog.