8 signs your attribution model is stifling your growth

Signs your marketing attribution model is stifling your growth

They’re waiting for you.

Droves of ripe, ideal customers who would totally prefer your USP over those of your competitors—if you only gave them the choice.

But for a variety of reasons, your marketing budget won’t allow you to reach them. You’re unknowingly ignoring these prime prospects, and not necessarily because the budget is tight.

No, often the far bigger problem is that too much of your advertising budget is tied up in channels that aren’t delivering the ROI that you think they are.

Today, a smorgasbord of offline and online advertising tactics deserve credit for nudging along each individual sale. But even though most every aspect of these tools is precisely measurable (and often in real time), painstakingly attributing varying degrees of credit to each “touch” is easier said than done.

Now add to this headache the fact that you’ll often have internal marketing teams and multiple siloed agencies clamoring for additional spend on their respective specializations. How can you defrost the foggy windshield of ego-driven metric massaging that often goes along with that? (Asking for a friend.)

Honestly, it’s enough to give any marketer an attribution aneurysm.

As a result, many marketers are unwittingly misallocating alarming portions of their advertising budgets. But it doesn’t have to be that way.

The gap between your current growth rate and the one you want to achieve gets a lot smaller once you’ve developed a sophisticated multi-touch attribution model. And a growing number of marketers are keenly aware of its importance.

According to a 2017 survey from the Interactive Advertising Bureau and Winterberry Group, the share of marketers planning to prioritize cross-channel measurement and attribution rose by 71 percent from the proportion that actually prioritized it the year prior.

But before you can take critical steps to earnestly optimize your attribution strategies and, ipso facto, the way you spend each marketing dollar, you need to be mindful of these eight telltale signs that your fundamental approach to attribution might be flawed from the outset.

Below you’ll find a condensed list of the most common attribution-related errors we see marketers making these days, so that you can know where to look for hidden inefficiencies—and unshackle your growth ASAP.


1. You don’t have an attribution strategy

We couldn’t create this list of mistakes and not start with the most glaring one of them all. But according to the same survey cited above, 43 percent of marketers still don’t prioritize cross-channel measurement and attribution. And in our experience, there’s usually one recurring reason for it.

Marketers today are expected to do more with less, and they’re expected to do it fast.

CMOs want XX percent more sales (or leads) than last year, and there’s nothing inherently wrong with that.

But that pressure can cause marketers to hit the ground running and try to beat competitors to the punch, without first taking a critical pause to ask, 'Wait, what is this investment actually giving us?'

As a result, marketers spray money everywhere. And once they start, the idea of pausing individual tactics to measure each one’s relative impact—we’ll talk about the smartest ways to do that in a moment—would temporarily hinder their growth and prevent them from meeting the boss’ expectations.

And so, they keep the pedal to the metal, never fully figuring out which combination of channels and tactics—at which specific investment levels—would give them the most bang for their marketing buck.

In those cases, once it finally does become apparent that they need an attribution model, it’s clearly an afterthought. And afterthought attribution modeling gets afterthought attribution results.

Because at that point, the process can feel a bit like untangling an attic full of Christmas lights over Thanksgiving weekend. It can certainly be done, but you’ll hate yourself and most of your family by the time you finish.

The moral of the story is, without an attribution model, testing is nearly pointless, because you’re not calculating the relative impact that varying investments within each channel are having on your success. You’re flying blind, and you can’t have real confidence in the budgetary dials you twist as a result.

In other words, you’re turning what could be a smart ensemble of hyper-measurable direct-response media… into a de facto branding effort.


2. You’re relying on last-click attribution

If you’ve read much on attribution best practices or listened to a marketing podcast, like, ever… this one should also be pretty obvious. But it still bears repeating, and here’s why.

Assigning sole credit to the ad on which a new customer most recently clicked (or viewed) before converting is a symptom of a much deeper problem: You’re not looking at attribution holistically.

Equally important as determining where that last click is taking place… is determining how consumers are getting to that last click.

And while last-click attribution is certainly much simpler and easier to measure than most other strategies, it presents some serious dangers for marketers.

In reality, some tactics are just by nature designed to optimize performance at different points along a customer’s purchase path: awareness, interest, consideration and, finally, purchase.

For example, branded paid search is hugely effective on prospects who are already at the bottom of the sales funnel. But if you pull too much budget to branded search and away from other tactics that are more geared toward building awareness or interest, before you know it, your funnel will be dry as a bone.

For that reason, it’s critical that you understand all the different touchpoints that contribute to a sale and grasp the integral role each channel plays along your customer’s path to a purchase.


3. You’re trying to measure the unmeasurable

Let’s be honest. In some rare cases, your marketing investments are just pricey backstage passes for your company’s leadership team. Let me explain.

One good example of this was highlighted in a recent blog from author and digital marketing evangelist Avinash Kaushik. In it, he paints a scenario in which your brand decides to sponsor a leading golfer. In exchange for a $22 million endorsement check, that golfer will now wear a hat with your brand’s logo on it every time he plays.

Naturally, your bosses are curious to know the impact of that investment. So you, the marketer, set out to find the answer.

After assessing the gross rating points for televised golf tournaments, conducting extensive surveys of consumers and/or CEOs at prospective B2B clients, and utilizing media-mix modeling and a host of other measurement efforts over the course of many months, you realize something:

Your CEO really likes golf and wants to rub elbows with top players in the hospitality tent at tournaments.

The first takeaway is, look, let’s please just call these investments what they really are, and stop wasting valuable resources trying to calculate the ROI they’re producing.

You’ve got enough to worry about in measuring tactics that really are measurable—and those demand your undivided attention.

Secondly, we won’t always suggest that you axe these investments entirely—sometimes they provide helpful air cover and brand recognition that can augment your more easily attributable efforts. But as true-blue direct-response marketers, we certainly wouldn’t deter you from at least having those conversations.


4. You’re relying only on directly attributable response vehicles

In the direct mail context, even though your letter might contain a campaign-specific 800-number, landing page URL and/or promo code, relying on those elements exclusively to attribute sales to that mailing effort would give you a grossly restricted view of its true impact. Here’s why.

Oftentimes, the recipient won’t choose to call the phone number that appears on your letter, nor will they take the time to type in the slightly longer-ish URL that’s listed. Instead, they’ll do a Google search for your brand or simply type in your brand’s homepage URL directly, navigate your site and, ideally, convert—perhaps even forgetting to enter the mail piece’s promo code on your shopping cart page.

As a result, direct mail gets no credit for this sale. Sounds a little silly, right?

In our experience, limiting your attribution to only directly attributable response vehicles will reduce a channel’s perceived impact by 50 to 70 percent compared with reality. In turn, marketers with this incomplete view will often stop investing in a channel like direct mail, cutting off at the knees what would soon become a potent acquisition tool—and severely throttling your scaling potential.


5. You’re not running match-back analysis

One critical step beyond measuring directly attributable response mechanisms is running what’s called a match-back analysis.

Basically, you take your file of new sales over a specific window of time, and then match them back to the list of people you mailed during a relevant period leading up to those sales.

Often, this match-back analysis will reveal that your direct mail campaign’s real impact was significantly greater than what your directly attributable measurement vehicles had suggested.

For example, let’s look at the initial direct mail test we ran for a large debt-settlement company.

A couple months after our letters hit mailboxes, the client expressed concern that the channel wasn’t performing up to their expectations. This was not at all an unusual fear among first-time mailers.

We explained to the client that they were seeing only the results from directly attributable response vehicles, and that after we run a match-back, based on our experience in similar verticals, the client could expect to see mail’s impact being 2-3x greater than what the surface-level results were showing.

Sure enough, the match-back revealed just that.

Today, the client uses direct mail to onboard over $1 billion in new debt every year.

There is, however, an even more precise way to measure the relative impact of direct mail (or any individual touch for that matter). We’ll discuss that in a moment.


6. Your lookback windows are too short (or too long)

As we alluded to in the previous section, you can’t accurately measure a touch’s true impact if you don’t understand the typical window of time during which the touch can have a lingering effect that leads to a purchase. This period is also known as a lookback window.

In the digital marketing landscape, the lookback windows are relatively short. A display ad, for example, shouldn’t be allowed a lookback window of more than a few days. Even that can be approaching eternity status, but it really just depends on the nature of the product and the amount of consideration that typically goes into making a purchase.

(It also depends on whether the prospect clicked on the ad or was merely served one—which would still trigger a viewability tag, but there’s no guarantee that the prospect actually saw the ad before converting.)

On the direct mail side, however, things are different. Mail is tangible. It’s physical.

It’s not at all uncommon for a letter to remain on a kitchen counter (or for its affixed card to get pinned to a fridge with a magnet) for months before the prospect finally converts.

In our experience, the typical lookback window for a direct mail piece is generally between six and 16 weeks—again, depending on the nature of the product.


7. You’re not running holdout tests

An even more precise way to measure the impact of a single touch is to run what are called holdout tests.

With this strategy, you’re taking a group of prospects and serving them a variety of ads across a variety of channels. At the same time, you’re randomly selecting a portion of that audience and serving them those same ads across those same channels, only with this group, you’re choosing one tactic to not employ in your engagement with these folks—we call them the “holdout” group.

By comparing the overall sales rate of the larger group with the sales rate of the smaller (but still statistically valid) holdout group, you’re able to see the isolated incremental lift you enjoy by employing the tactic you didn’t use on the holdout audience.

Now, if you’re running holdout tests among strictly online channels, these analyses can prove quite difficult, especially now that Google will no longer share user IDs externally.

In online environments, you’re not starting with a completely positive, deterministic understanding of a prospect’s identity, so you can’t guarantee that the person you touched in one channel is the same exact person you touched in another channel.

One way you can help solve for this is by using geotargeting. For example, let’s say you run paid search across the entire country. Then, in one state, you’d add just Facebook ads. In another state, you’d add just display ads. In another state, you’d add both Facebook and display, etc. By comparing the differences in performance among those geographic areas, you can start to get a feel for the incremental lift that each individual touch provides.

While holdouts among digital media are slightly more complex, direct mail is relatively straightforward.

You might never fully eliminate the debate over the degree to which each digital touch contributed to a sale, but with direct mail, the holdout is irrefutable. Employ all the other tactics you want, including direct mail, then select a portion of the audience to not mail, and the difference in performance is simple to measure and clear as day.

Mail’s impact is truly incremental, and justification for future investment is bulletproof.

Also, in case you’re doubting our suggested lookback window for direct mail, a holdout test is where you’ll really see concrete proof of mail’s uniquely long-lingering impact. Barring any freaky external forces, the holdout group should, for a brief time after mailing, perform at a sales clip that’s nearly identical to that of your mailed group. Then the mailed group will begin to show a noticeable relative spike in sales rate that will last, again, typically up to 16 weeks.

Once the holdout group and the mailed group start performing at the same clip again, it’s a pretty good indicator that your direct mail sales curve is wrapping up, and that’ll give you the best idea of what your direct mail lookback window should be moving forward.


8. You’re not effectively onboarding offline data in online environments

Another way to make your digital attribution model even more precise is to start with rich, terrestrial direct mail data and onboard those offline identifiers in online advertising spaces.

This way, you’re not starting from a place of total anonymity in your online audience targeting.

You can deliver consistent, synchronized messages to prospects across multiple media and isolate the incremental lift that each of your digital tactics is providing to your direct mail performance.

For example, some of your prospects would receive direct mail only, some might receive direct mail and email, some might receive direct mail and Facebook ads, some might receive direct mail and Instagram ads, some might receive direct mail and email and Facebook ads, some might receive direct mail and email and Instagram ads, some might receive direct mail and email and Facebook ads and Instagram ads, some might receive… you get it.

These offline-to-online integration holdouts are proving especially fruitful in the multichannel campaigns we’re running for a leading credit-optimization service. Currently, integrated Facebook ads are boosting the client’s direct mail sales rate by 14 percent on average, and by as much as 38 percent within the best-performing offline file.



As you can see, there are numerous ways a brand can commit costly attribution errors. If you’re not careful, you might wind up misdirecting huge segments of your marketing budget to channels that aren’t maximizing growth for your brand.

As a result, you could be unknowingly ignoring thousands of ideal customers every month, simply because your precious budget is tied up in bottom-of-funnel (and ultimately limited) marketing tactics.

We agree, at times, dialing in your attribution strategies might seem like much more of an art than a science. But by avoiding the all-too-common pitfalls mentioned above, which can make attribution modeling far more nebulous than it needs to be, you’ll be giving yourself a distinct advantage over the nearly 50 percent of marketers who still don’t prioritize this most critical endeavor.


Want help refining your multi-touch attribution model, so that you can quickly start getting more out of your marketing budget?