PPC & Media

Valuing View-Through Conversions

View-through conversions (also called post-impression conversions) are when a person is exposed to a display ad, does not click, but goes on later to visit your website and convert.

In the world of search marketing, this is a somewhat foreign concept, as search marketers tend to live and die by the click. Search, by its very nature, has a pre-qualified audience based on their intent to seek out information. Clicks from search ads connect your brand with a sought after need.

Display advertising, on the other hand, is much more driven by influence. As a marketer, the best we can do is reach the right type of person, preferably in the time and location where they’re most likely to engage with our brand. Even when all those factors fall into place, a person is still only connected when an impression is served and your target is compelled by what they see to take some action. In display advertising, the impression still reigns king.

How do we value a view-through conversion?

In concept, we can accept that impressions are influential. But for years we, search marketers, have been held to strict ROI and CPA goals, with every action having a clearly defined cost and value. We have been conditioned. So what are we supposed to believe when we see numbers like this?

View Through Conversions

Amazing, or smoke and mirrors?

Unfortunately, numbers that drastic are difficult to accept, and are often dismissed entirely. Giving no credit to a view-through conversion is not fair, though, as there have been countless studies to show that lift and influence are very real.

We also can’t rightly give VT conversions 100% value. Especially when utilizing remarketing (compared to other forms of display targeting), many of your ads will undoubtedly serve to consumers who would have already converted without the added influence. Ads will also serve impressions that go un-seen (below the fold), but in the eyes of the algorithms a conversion is still valid. These facts make it abundantly clear why VT conversions are so inflated.

So what’s the true value? Somewhere in between.

Finding your proper value

Unfortunately there’s no cookie cutter answer that can be applied to every campaign, but we can perform some tests to find out what is right for you.

Reduce your lookback window

VT Lookback WindowLogic tells us that the longer the gap between an ad being served and an action being taken, the less likely the ad actually influenced a decision. Most display settings (Google or otherwise) naturally default to a 30 day lookback window! Drop your lookback window to 1 or 2 days. This will give you a much clearer picture of ads that actually influence behavior.

 

Once your settings are properly configured, the real analysis can begin.

The PSA Test

A PSA Test (Public Service Announcement) splits your audience into a control (PSA) and test (Branded ads) to compare how many conversions occur across each. The control group, a PSA ad, will have no influence on your audience. It also will likely continue to convert, based on the criteria we described before. Your normal Branded ads will also continue to convert, likely at an even higher rate as you are reminding visitors of your brand and guiding them to take action. That ratio will be our modifier to properly weight view-throughs.

Banner Ad Comparison

  1. Segment your audience to create two equal lists that do not overlap.
  2. Your first campaign will use your Branded ads, while the second campaign will use PSA control ads.
  3. Allow your ads to run for a sufficient amount of time to collect enough data. Remember, we’ve recommended a 1-2 day lookback window, but if you’re still set to 7 days or more, you’ll have to run this test for a minimum of a week to allow for proper collection.
  4. Normalize your data and compare results:

VT Conversion Formula

Let’s take a look with some real numbers to better understand

 

ImpressionsVTC
Brand Campaign 100,000 100
PSA Campaign 80,000 72

 

To normalize the data, we multiply the PSA campaign results by 1.25, leaving us with:

 

ImpressionsVTC
Brand Campaign 100,000 100
PSA Campaign 100,000 90

 

Weighted VTC = (100 – 90) / 100 = 10%

What this tells us is that only 1 in 10 view-through conversions actually influenced someone into taking action. So is that good? Well looking at our results from before, we had nearly 10x as many view-through conversions as we did click-through conversions, so even at just 10% weight, this would actually still be pretty significant. Your percentages will vary from this example, of course, based on your audience, settings, and creatives. With such a test you can finally get a real understanding of just how effective your display ads truly are.

Craig Galyon

Craig Galyon - Digital Media Team Manager

Craig Galyon (pronounced "Craig") is the Digital Media Team Manager at SwellPath, with a background in PPC, Display, Rich Media, and Social advertising. When he's not managing killer media strategies for his clients, you'll find Craig brewing beer and playing competitive kickball. Rarely together.

6 Responses to “Valuing View-Through Conversions”

  1. Henrik

    Brilliant and so simple, that I’m embarrassed we didn’t think of it ourselves. How about the PSA’s, do we create them or does Google offer PSA’s?

    Reply
  2. Henrik

    One more question if I may: What url do we use in a PSA? Aren’t we restricted to link to our own services/domain? As far as I know Google has to approve all ads.

    Reply
    • Craig Galyon

      Thanks for the questions Henrik!

      1) If you’re doing Remarketing through Google, then unfortunately the test will have to come from your own budget. If that’s the case I’d usually set a small percentage of impressions (10%-20%) for the test, so you can run it for a couple weeks without eating up too much of your spend. Some 3rd party companies, like AdRoll, will run the test for you on their dime, but only if you’re spending a significant amount.

      2) Go ahead and link to your website. The purpose is to see if people will be impacted by views, not clicks, so it shouldn’t matter what URL the click leads to. The idea is that they won’t be clicking anyway.

      Reply
      • Henrik

        1) Money is not the problem ;-) To clarify: Do we design our own PSA and if so is it really accepted by Google that we link to our Domain. Let’s say we design a Red Cross donation ad and link it to our website. Would that really be ok with Google or am I missing something

        Reply
        • Jackie

          Henrik,

          I believe that PSA testing is a feature in adwords so I assume you setup the test and google automatically will place their PSA google brand ad in the control portion of the campaign.

          This is great stuff, have been trying to figure out how to do this forever, thanks for the awesome content Craig!

          Reply
  3. Veronica

    How can we make sure that the segment of audience that was served PSA ad actually WON’T see the Branded ad?

    Reply

Leave a Reply

Get Our Newsletter

To keep up with the latest news from SwellPath on digital advertising, analytics, and SEO, sign up for our newsletter.

Archive