How to check if display advertising is driving incremental sales? (with case study)

In 2016, digital display ad spending will eclipse search ad spending in the US for the first time [Source: Emarketer]. Overall, one in five dollars devoted to digital in 2016 will go to “banners and other” digital display ad types. Budget is shifting to display because the technology purportedly targets intent and is a superior method to content that drives click-throughs. Some studies (download PDF report from Nielsen ) show that click activity has little to no causal relationship to sales conversions.

Click-through rate is not the right metric to measure brand impact – virtually no relationship exists between clicks and brand metrics or offline sales.

The challenge with display advertising is getting a read on exactly what it delivers to the business.

Why post-view conversions on display advertising are an imperfect metric?

Display vendors (Rubicon, Quantcast, Criteo, AOL, Youtube, etc) provide a report on post-view conversions. On its own this is prone to value-inflation for the following reasons.

  1. The time to convert between ad view and sale is too long
  2. Overlap with other marketing activities makes it difficult to separate true impact
  3. Post view conversions is a greedy method
  4. Not all the claimed conversions can be attributed to the display advertising.

I will explain with the graphic below. It’s taken from one of our clients (sensitized data is presented). A little bit of background. This team has done a fantastic job nurturing a customer-base. There is a very strong loyalty for the brand and the email newsletters – which are refreshed with new content every week. Brilliant stuff!

So, when they tried out display – we did the audit. Here’s the view out of Polytab attribution portal . What the attribution was telling us in overlaying the ad views on the shopping journeys was that the touches were vastly out of sync with the actual purchase behavior – and it was email (Silverpop) that was the true revenue driver. That was our recommendation – that we did not feel display influenced sales.


To show you where the metrics come from I will share the raw data under the hood. The screenshot of the spreadsheet below is the activity for one shopper across all his touch points with our client.

As you can see the ad was served on Nov-16. The display vendor claimed three orders (highlighted in green below). Even a cursory glance told me that the sale was best attributed to the email that went out on Nov-30. Furthermore, even the most optimistic scenario could not claim three orders to the single ad view two weeks ago. This was a good representation of the post-view conversions.

So we went back to our client and shared the facts. They are now revisiting the strategy.

Sample data on post-view conversions claimed by display vendor
Sample data on post-view conversions claimed by display vendor

It’s not gloom and doom. The technology assists sales. But here’s how to measure the incremental sales delivered through display advertising.

How to measure if display advertising delivers value?

Post-view conversions are a healthy metric. Most vendors should provide the list of orders they influenced – and if they do not – you should reconsider that relationship. But augment that list with the following enhancements

  1. Measure Prospected vs Remarketed sales: Arguably, the biggest benefit for going with a display network is they provide a pool of look-alike audience that can be targeted to broaden the retailer’s target market. Break out the post-view conversions across the prospected and the remarketed audience.
  2. Measure shoppers (not orders): As with the example above, you are going to get repeat purchasers. The impact of display advertising is most on the first time purchase post-ad
  3. Check shopper interest before vs after the ad view: The hypothesis test should assess if shopper behavior is the same before/after the ad view. The test we did for the client above showed that the shopper behavior in the 12 hours before the ad view was statistically identical to the behavior 12 hours after. The test showed up the same results by extending time windows to up to 72 hours.

Why Polytab attribution works

Polytab is Infernotions’ attribution offering that guides retailers on how to grow sustainably without blowing the marketing budget. Most importantly it’s a turnkey solution that is priced on a sliding scale to adjust around retailer traffic. It also delivers results via a beautiful interactive display.

One of the advanced packages is attribution to display advertising. If display advertising, or attribution in general is important to you, check out the demo by clicking on the link below or give me a call at the number on our web site.

sign in as guest

Image of old broken Tv by schmilblick reused without changes with permission under creative commons.

Leave a Reply

Your email address will not be published. Required fields are marked *