Each “touchpoint” associated with each individual has numerous attributes or characteristics associated with it.

For example, each search impression or click has attributes such as keyword, engine, placement/ranking, ad creative, offer, landing page, day-part, and more.

Each display ad impression or click has a different set of characteristics associated with it and the same is true for each channel.

When all the data from all these touchpoints across all your channels has been assembled, not only is the attribution management process able to assign the appropriate amount of credit to each touchpoint (and each marketing effort), but a value-added process called touchpoint analysis can also take place.

This can be a treasure trove for search marketers.

What Can Touchpoints Tell Us?

By looking at the group of touchpoints associated with an individual’s “touchpoint stack,” and in turn all of the touchpoint stacks of all the individuals exposed to your marketing efforts who have and have not converted, a number of findings emerge.

In particular, a picture of your various conversion funnels (for each of your products, product lines, products types, etc.) will come into focus.

You’ll be able to see the relationship between organic and paid search – and for that matter direct navigation – and the extent to which each plays a role at each stage in the engagement/sales funnel. You’ll learn where in the funnel both branded and non-branded keywords play a role.

You’ll identify the specific keywords that are particularly effective at generating top-of-funnel activity (“openers”), the keywords that appear at the middle of the funnel (“advancers”), and which perform best at the bottom of your funnel (“closers”).

Again, these findings may vary widely for each product or service you sell, so there may be many different funnels you will uncover with these insights.

You’ll also be able to identify the sequence of touchpoint attributes that generate the best performance toward your marketing goals, as well as any “typical paths” through the conversion funnel.

For example, do people start with a specific non-branded keyword in paid search at top-of-funnel, move to a specific branded organic keyword mid-funnel, and direct navigate to your site at bottom-of-funnel?

Do certain keywords used at one stage of the funnel lead to specific keywords at the next stage of the funnel?

Turning These Insights Into Optimization

Once you’re armed with these types of touchpoint analysis findings, you can focus on optimization strategies and tactics to utilize the most productive touchpoint attributes and apply more of your budget on the best performing combination of attributes.

For example, is a particular keyword, ad creative, bid price, landing page and day-part combination found to be the best “opener” for a specific product type?

If so, decrease budget on less productive attribute combinations and devote more budget to this more effective combination. Of course, it’s also important to make sure that the combinations on which you decrease budget aren’t top performers at other stages of the funnel.

You also have the opportunity to strengthen the attributes associated with each funnel stage.  If a particular keyword, engine or ad creative is an effective opener, you can modify the copy of the landing page so that it doesn’t try to “close” the conversion – but instead helps cultivate the prospect.

If you find that there is a particular sequence of keywords that’s more effective than others, perhaps you can embed a call to action using that keyword sequence on the landing page used for the previous stage of the funnel.

As a search marketer, your very nature is to mine insights like these and convert them into optimization strategies that you can test and refine. If you or your colleagues are already doing attribution management, consider performing touchpoint analysis in order to uncover this additional dimension of valuable insights.

Among these more ancillary factors is “frequency” — specifically the frequency with which online users are exposed to a given marketing tactic in advance of an eventual conversion.

A byproduct of applying multidimensional attribution modeling to your cross-channel marketing results will be findings that uncover the number of impressions generated within one channel beyond which an acceptable return in conversions fails to take place via other channels.

Of Course Display Drives Searches

Let’s take the example of conversions that are recorded via the search marketing channel. Not only do we all intuitively know that online display ad impressions drive users to perform searches for specific keywords that eventually lead to conversions that take place on the search engines, but through the application of attribution management we can actually quantify this impact.

But one dimension to those attribution findings that is often overlooked — that also needs to be quantified — is the maximum number of display impressions for which you are willing to pay to produce a search conversion. In effect, you need to start thinking about establishing cross-channel frequency caps.

Cross-Channel Frequency Caps

Digital media planners and buyers are familiar with applying frequency caps to their display assets based on some number of impressions above which the ads have historically failed to produce an acceptable return based on conversions recorded through clicking directly on the ads.

But the reality is that a frequency cap established using that one-dimensional criteria might be limiting a very positive return that those ads may be generating via conversions they drive within the search channel.

Once this frequency data is revealed through your attribution results, this number will probably differ from publisher to publisher.

But the beauty of multidimensional attribution is that you can easily drill down to as granular a level as you wish in order to uncover these findings, which will inform the actions you take (across the combination of the publisher, keyword and search engine dimensions, for example) to optimize the performance produced by the interaction of the display and search channels.

Don’t Leave Affordable Conversions On The Table

As an example, imagine that the acceptable return on advertising spend (ROAS) for a given marketer is 8.5. In this case this ROAS includes both the cost of the search clicks, as well as the cost of the display impressions.

In the chart below, you’ll see that running a specific paid search ad on a specific search engine produces an ROAS of 8.5 or greater until online users have been exposed to eight display ad impressions recorded by a specific display ad on a specific publisher.

After that point, the ROAS drops well below 8.5 as additional costs are incurred but conversion delivery tapers off. So in this case marketers would want to apply a frequency cap of eight to this particular display ad on this particular publisher.

The fact is that in this example, the frequency cap the marketer probably would have applied based solely on the ROAS of conversions recorded through clicks directly on the display ad - would have been five, leaving lots of affordable conversions that could have been recorded via the search channel still on the table

Now the impact of such a frequency cap on the other dimensions of your marketing ecosystem should certainly be assessed to ensure that performance gains achieved for this keyword/search engine/publisher combination do not have corresponding losses elsewhere.

By looking at the attribution results for display ad frequency across all keywords and engines, odds are you’ll be able to identify a number of instances where significant media efficiencies and performance gains can be achieved by toggling up or down impression frequency caps as needed.

It does this by scientifically calculating the impact that every marketing touchpoint experienced by your prospects has on achieving a specific marketing goal – such as conversions, revenue, return on advertising spend (ROAS), etc.

But where the attribution process ends for most marketers is with the acquisition of a new customer and looking solely at media metrics – failing to extend the attribution exercise to include the longer term enterprise value or quality of that newly acquired customer.

Look At Value Beyond The Initial Transaction

Suppose that after performing attribution, it was revealed that a given search engine actually produced 1,000 conversions at a given ROAS for a specific keyword for the initial transaction that it produced.

Meanwhile, another search engine produced just 300 conversions for the same keyword at a given ROAS on that initial transaction.

If your analysis stopped there, it would be pretty simple math to determine which search engine to channel future budget to for that given keyword (assuming sufficient search inventory existed for either option).

But if this attribution exercise included looking at the some “customer quality metrics” that were revealed over time about the 1,000 customers compared to the 300 customers, very different investment decisions may result.

What if you went a step further and looked at different metrics associated with those 1,000 customers and those 300 customers at three months, six months and 12 months after that initial transaction?

For example, the following metrics might be very telling about the quality of your customer:

• How many have made subsequent purchases and how many were one-shot wonders?
• How many are still active subscribers and how many has attrition eliminated – requiring you to now have to spend acquisition dollars to “re-activate?”
• How many have bought high-value (to you) products?
• How much revenue have they produced and what is their predicted revenue contribution within an appropriate customer segment?
• What is their average lifetime value (LTV) taking into account profitability or influence from a social or viral perspective?

After analyzing these additional metrics the results may surprise you, as it’s entirely possible the search engine that produced only 300 conversions might have resulted in more quality customers over time.

Feed Customer Quality Metrics Back Into Attribution

The example above obviously extends beyond search to online display ads or any other media channel – as well as across various channels – but the point is that once the quality metrics above have been collected, they can be fed back into your attribution engine.

Then they can be used as a success attribute to predict and ultimately optimize future marketing performance based on not just the value of the first transaction and your ability to achieve your acquisition goals, but to optimize for long term ROAS, ROI and enterprise Lifetime value (LTV).

This is a great start at taking the attribution process to the next level, but you owe it to yourself to not stop here. After all, at this point you’re only looking at two dimensions of your success: the media metrics associated with producing a new customer and the media metrics associated with producing the highest quality customers.

What’s next is adding demographic/audience attributes of those customers to the attribution process.

What Else Produces Highest Quality Customers?

By either collecting age, income, geography, gender, and other demographic data directly from your customers, or by obtaining it from third parties like Exelate, BlueKai, Acxiom or Experian and feeding it into the attribution process, you can not only produce insights on how to optimize for the media tactics that produce the highest quality customers, but use those media tactics to target prospects with the demographic traits associated with your highest quality customers.

Attribution also provides insight into the media consumption behavior of the desired audience by providing insight into content by publisher consumption.

But the key is attribution. The multi-dimensional nature of attribution enables you to analyze your marketing success across an unlimited number of stimuli, response metrics and customer attributes to produce a scientifically valid set of metrics for what truly contributed to your success across your entire marketing portfolio.

By adding the dimensions of long-term customer quality, as well as customer demographic attributes, you truly elevate your attribution efforts to best-of-breed status.

So not only can a marketer learn which publisher is producing the best ROI or most revenue, but he/she can learn which creative within that publisher is producing the best performance based on the offer, size, placement, day, time and frequency (among other campaign attributes) – taking into account all the influences that exist between campaigns and channels.

As long as there is available inventory, the marketer can then devote budget to the best performing combination of campaign attributes in addition to now being able to find inventories with similar profiles.

How Does This Apply To Search?

One of the very basic findings that attribution uncovers within search marketing results are the keywords, ad creatives, and landing pages that serve as the best “introducers,” “influencers” and “closers” within consumers conversion funnel.

In simple terms, attribution does this by analyzing all touchpoints that were experienced by its target audience – those on which a conversion took place and those that took place earlier in the funnel on which a conversion did not take place.

It then calculates the impact that those non-converting touchpoints had on an eventual conversion by comparing conversions on which those earlier touchpoints existed to conversions where those earlier touchpoints did not exist.

No search marketer will be surprised to hear that one type of insight that this attribution process uncovers is that non-branded keywords searched early in the conversion funnel don’t get as much credit as they deserve for an eventual conversion that takes place on a branded keyword. Without being introduced to a brand and influenced by a brand due to certain keywords, a branded search/conversion may never have taken place.

Attribution quantifies that credit and in doing so identifies in a multi-dimensional way the combination of search marketing attributes that serve as the best “introducers,” “influencers” and “closers.”

Applying Those Findings To Other Channels

Once a marketer has discovered the keyword, ad creative and landing page copy that serves as the best “introducer,” these attributes can be translated into the channels being used to introduce new prospects to his/her products.

This could mean the insertion of a keyword that didn’t previously exist in a print ad, TV ad or email campaign into new versions of those campaigns/channels. It could mean changing the messaging, call to action or offer of those channels.

Similarly, if you knew the combination of keyword, creative, offer, message, etc., that served as the best “closer,” these attributes could be translated into use within the channels that marketers choose to use as their best conversions.

Delving A Little Deeper

Taken one step further – which once again, a beauty of attribution management is its multi-dimensions– these best performing “introducer,” “influencer” and “closer” attribution combinations could very well vary from one product to another, one business unit to another, or any other business-specific criteria.

Armed with this level of granularity of findings, marketers could create variations of offer, creative, calls to action and messaging for each channel for each product, business unit, etc.

Will every marketer’s post-attribution search marketing findings point to dozens or hundreds of combinations of creatives that need to be crafted? Probably not, but if marketers discover that their best performing keywords, offers and messaging within the search channel are nowhere to be found within their other channels, there is definitely some low-hanging fruit waiting to be harvested.

For a while now, companies like Simpli.fi, Magnetic and Chango have been offering marketers the ability to identify Internet users who have performed a search on a given search term on one of the websites with whom these vendors have partnerships.

Marketers then use that information to target these individuals with online display ads via the ad exchanges.

The search engines have actually been doing this between their paid search platforms and their own display networks for a while as well. It’s a very cool concept and has proven to be particularly effective under certain conditions.

Enter Attribution Management

What elevates this cool concept to the super-cool level is the addition of attribution management to the process.

As a bit of background, the attribution process provides marketers the ability to identify and monetarily quantify the ad exchanges, adservers, publishers, creatives, sizes, offers, and any other attributes associated with display ads that produce the best return on investment toward any of the marketer’s specific business objectives/metrics.

It is also able to identify and monetarily quantify the affinity between specific display ads/variations and the keyword or keywords that were subsequently searched by users after exposure to those ads – including those keywords that serve as the most effective “introducers,” “influencers” and “converters.”

Retargeting on Steroids

Armed with a knowledge of the ads/placements that produce the most bang for the buck, as well as the knowledge of what keywords have an affinity with which display ads, marketers who want to pursue more effective search retargeting can do the following:

• Identify the keywords with which the most effective display ads have an affinity and retarget those individuals who performed searches on those keywords with those most-effective ads/placements.
• Identify the keywords that serve as the most effective “introducers,” “influencers” and “converters,” and retarget individuals who performed searches on these terms with the display ads/placements with which they are associated.

As an example, people who are exposed to highly-effective display ad “A” have an affinity to eventually search for “Bermuda vacation rental.”

With this knowledge, marketers can use a search retargeting service to find all those people who searched “Bermuda vacation rental” and serve them display ad “A” via the ad exchanges.

Conversely, marketers who know that the keyword “digital camera” is their best “introducer,” that “under $300 digital camera” is their “best influencer,” and that Nikon Rebel XT price” is their best converter, can identify all those people who searched for those terms and serve them display ad “B,” “G,” and “M,” respectively based on the affinity of those keywords with those ad variations.

Measure, Test & Refine – At No Additional Cost

Obviously, attribution, search marketing and search retargeting are all iterative processes, and not one-time projects. So it is imperative to continue to monitor the results of all three and tune them accordingly as new insights are discovered about performance and affinities.

The great news is that this can all typically be funded through reallocation of currently ineffective spending that’s discovered through the attribution management process.

Now, imagine a room with just three men or just three women. Were we given the same task we could still carry it out, but not only would our selection be far more limited than when we could select from 20 people, but we might not find any of the men or women to be truly attractive.

We might have to “settle” on the most attractive person. In reality, with a greater selection of men or women comes a greater likelihood that we’ll find someone truly attractive.

A Marketing Beauty Contest

With marketing attribution management, the same basic concept holds true. If the only marketing performance data sources used to fuel your attribution engine are from one search engine and only online display publisher buys, when it comes to attributing credit for your conversions there will be a limited number of marketing touchpoints available for use in the attribution equation.

On the other hand, you can imagine that the attribution process has a lot more to consider when identifying how to distribute credit for your conversions when you need to include performance data from a variety of sources such as the following:

• several  different paid search programs
• organic search from multiple engines
• publisher buys
• rich media
• video
• ad exchanges
• any other display ad sources not captured by your ad server
• Omniture data
• data from your email tool
• social media sources
• your direct mail results
• summary data obtained from your TV, radio and print campaigns

It also has a much better chance of identifying the channels, campaigns, and campaign attributes (size, placement, publisher, keyword, timing, creative etc.) that truly impact the eventual conversion.

In effect, it has a better chance of finding a “person” that’s truly attractive. And of course any number of data sources that fall between these two extremes still has a better chance of producing a more accurate picture of where credit for your marketing success should be given.

So Does Search Suffer By Comparison?

First “Yes,” and then “No.”

In general, we’ve found that as more data sources are used in the sophisticated algorithms that calculate the true impact of each marketing touchpoint on conversions, the less credit that search gets for those conversions.

Just think about it: if you had 20 data sources being used in the attribution equation – even if ten of them were responsible for only a fraction of one conversion each, that would still eat away at the amount of credit that search would get for your overall universe of conversions.

Unless you include marketing performance data from those sources, you’ll never know what, if any, impact they had on your overall marketing success, and how much direct conversion credit they steal from search.

That said, we’ve seen that the more data sources you use in the attribution equation, the more correlations and synergies that marketers and their agencies are able to identify between search tactics and tactics used in other channels.

• Which display ad publishers, creatives, sizes and placements drive the most profitable conversions taking place on search?
• Which display ads or sequences of ads produce searches for which keywords or sequences of keywords?
• What size spikes in conversions are produced by which TV or radio ads in which markets at which timeslots?
• Which print ad or direct mail creatives and timing impact conversions via search?

The answers to these questions typically do not result in a fundamental decrease in search investment, but simply in a reallocation of overall investment to the combination channels, campaigns, and campaign attributes that help produce the most conversions across all channels – but particularly via search.

The Takeaway

When it comes to producing the most accurate attribution possible, the more channels the better. Start with search and whatever channel you intuitively feel has the greatest impact on conversions at your organization, then add additional channels as bandwidth and your growing mastery of the attribution process allows. As you do, the prettier things will get.

A lot can be said for the wisdom and fairness of such a compensation model given the ability of some attribution platforms to accurately and fairly calculate this credit. But gaining buy-in from all the disparate and competing constituencies affected by it would pose a significant challenge. Perhaps a topic for a future (much longer) column…

However, one area where this same idea makes an awful lot of sense, and is more easily implemented, is in how brands or agencies compensate their own staffs for the success of their efforts.

Attribution done correctly uncovers the true monetary contribution of each channel, campaign and tactic in your organization’s overall marketing success.

As a result, you may be forced to reassess how you compensate the people within your organization or risk failing to capitalize on the efficiency and effectiveness you could be extracting from your marketing spend and from the efforts of your marketing  teams.

Option #1 – Give Compensation Where It Is Due

As a simplistic example, let’s say your digital marketing organization is made up of three separate teams responsible for managing search marketing, display and email.

The performance of each team is measured by the conversions, revenue or ROAS their respective efforts produce – based on the traditional method of attributing 100% of credit for each conversion to the “last act” (often a click) that occurred prior to the conversion.  And the compensation and incentives of each team are tied to certain standards that are calculated by this same method.

But once attribution management is performed, and monetary credit for the conversion is spread across multiple touchpoints and channels, there will be certain channels/teams whose efforts have been undervalued by the traditional way of measuring their performance (such as display) and those whose efforts have been over-valued (such as search). At that point, your organization will have a decision to make.

Do you simply maintain the same standards for compensation, with one or more teams taking a hit in their compensation, and one or more teams getting an unexpected boost or use the output of attribution to create an environment that is more transparent and synergistic based?

As you can imagine, not only may the aforementioned employee revolt occur, but your teams will have no motivation or desire to work together towards the organization’s overall marketing success.

Option #2 – Create Shared Success Goals

In the example above, perhaps a less disruptive and more constructive plan might be to establish a shared set of performance metrics and compensation incentives that all three teams can work towards.

Establishing combined revenue, overall CPA, or total ROAS across all three channels – using your attribution metrics as the basis for this measurement – will motivate all three teams to pull from the same end of the rope. Channel-specific success will cease to be important or rewarded – only overall organizational success.

So when inter-channel affinities are identified by the attribution process – such as online display creative that drives searches on certain high-converting keywords, or the sequencing of certain creatives or channels that produces the highest return, or the maximum frequency of certain tactics before they cease to be effective, or the channels and creatives that serve as the most effective introducers/influencers/closers are discovered – everyone in your marketing organization has incentive to take advantage of those findings.

Optimization of the overall marketing performance will be the one and only objective and channel turf-wars will be a thing of the past. Mind you, in organizations that have multiple product lines and/or Business Units it could be extremely important for the CMO to understand how this complex engine is actually working. For example, is Business Unit 2 actually driving conversions for Business Unit 1?

Search marketing and search marketers have long served as change agents within larger marketing discipline.

Here’s an opportunity for them to lead the way in establishing a truly fair, truly team-oriented approach to organizational success, using attribution management as their enabler and their authoritative scorecard.

In fact, a company not pursuing these activities now is almost universally viewed to be at a disadvantage to its competitors who are so engaged.

Well, marketing history appears to be repeating itself in the form of attribution management. It’s clearly the “new search” in many ways.

The Wild, Wild West Of Attribution Management

Attribution management is the science of calculating the contribution that each marketing touch point experienced by a consumer has on generating a “conversion action” – such as a purchase. Once the true attribution of each touch point is calculated, more intelligent decisions can be made by marketers about investments in the specific channels, campaigns and advertisement attributes that produce the best return on investment.

Despite the amount of marketing industry ink devoted to it over the last couple years, attribution is still at its infancy in the same way search once was. A number of leading providers have arisen in this space, and yet each has a different approach to how attribution is performed. Lots of smaller players claim to “do it.”

Now, both interactive and full-service agencies alike are carefully eyeing attribution as a possible threat and as a possible new revenue stream at the same time. Some agencies are already considering the “do we build, outsource or acquire” decision. It’s the wild, wild West. Sound at all familiar?

Channel Early Adopters

Much like search marketing, there are early adopters of attribution – financial service and insurance companies, travel and hospitality providers, and major retail players. These are large B2C marketers that market themselves through many channels, have an abundance of data, many thousands of transactions, and lots to gain if they can move the needle even a few percentage points of efficiency across their entire market spend.

These companies also have the required amount of money, staff and risk tolerance to allow them to explore new channels and new disciplines with the possibility that they may fail. Smaller companies with much less to invest in “experimentation” remain cautious in wait-and-see mode while more progressive organizations prove or disprove attribution’s viability. Once again, we’ve tread this path before.

The Rise Of “New Marketers”

Job titles containing “search marketing,” – specialists, analysts, managers, directors, and VPs – were unheard of a decade ago, prior to the realization that specialists of this type were necessary to compete effectively within that new channel.

Today, a growing number of companies have recognized the need to sprinkle their marketing organizations with titles like marketing analyst, marketing engineer, or marketing analytics specialist – PhDs in statistics and other mathematical disciplines who apply “science” to the art of marketing. Professionals of this type are already helping guide their more qualitative marketing colleagues on the quantitative ramifications of their work as virtually every marketing channel becomes measurable.

These “new marketers,” validated by the results they produce working within isolated marketing channels such as search, online display, and database marketing, will help evangelize attribution management as they work with third party providers in this area to produce increases in performance across organization’s entire marketing ecosystem.

A New Way Of Thinking

The rise of the search channel forced marketers to think differently – “pulling” consumers to our websites who were already interested in our products and service – rather than “pushing” our marketing messages to consumers who may not yet be interested.

Perhaps most importantly, attribution management is also forcing marketers to think differently. Instead of defining the performance of marketing channels in silos, each producing its own ROI as calculated by the value of the conversions recorded through that specific channel, attribution is defining the performance of channels by their true contribution to organization’s overall marketing success.

Both the “points they score” and the “assists they provide” can now be considered when judging performance.

Editors Note: Manu Mathew is participating in a free Search Marketing Now Webinar, Marketing Attribution Demystified: Ask The Experts, at 1pm EDT on March 15, 2011 – Register Now!