Measurement Matters: Setting the Correct Lookback Window

A key input to campaign measurement is properly setting your attribution windows — or lookback windows, as they are often called. Simply put, a lookback window is the period of time you give yourself to “look back” at the ad exposures that contributed to a conversion. A 48-hour lookback window, for example, only considers a person’s ad engagement over the two days leading up to a conversion.

 

Most Lookback Windows are Too Short

We’ve done research on lookback windows and we often see cases where advertisers set lookback windows for a period of time that’s simply too short to understand a person’s actual purchase journey. Since small windows are more likely to give conversion credit to lower-funnel marketing activities like remarketing and search, advertisers who use them risk missing out on conversion credit for upper-funnel tactics that drive awareness and consideration.

And while it’s true that capturing any conversion — either as a result of an impression or a click — is considered a success metric in some campaigns, skewed measurement approaches can hurt you in the long run. What’s the risk? You may find yourself increasing investment on lower-funnel efforts like retargeting and search, to diminishing returns, while neglecting upper-funnel tactics. It’s these upper-funnel branding efforts that continue filling the funnel by driving the site visits that you subsequently retarget or generating the search queries against which you buy search ads. Failing to manage the full funnel ultimately leads to missed opportunities for your business.

 

Set Your Lookbacks With Confidence

Assumptions about the consumer behaviors can end up entirely defining how your campaign performance is measured. For this reason, it’s crucial to arm yourself with the right data and approach to setting lookback windows.

There are three common ways to find the proper lookback period for your product. Each method offers a different approach to defining the proper lookback window, and they have their unique pros and cons.

 

educated-guess

The Educated Guess

What is it? Informed knowledge based on historical information, customer research and general industry knowledge. Consider your products, price points, ad frequency, brand awareness and general consideration window with this approach.

What does it show you? Directional insights on the time it takes for your customers to make a purchase decision.

Note: The Educated Guess may work as a step 1, but it should be followed by quantitative method.

lift-test

The Lift Test

What is it? Lift is considered the gold standard for measuring ad effectiveness with causality. It’s an A/B test with randomized test (people who see ads) and control groups (people who don’t see ads).

What does it show you? Quantitative insights into the time it takes for the conversion rate of the Test group to converge with the Control group, informing the ideal lookback window.

Note: Lift testing is not possible on every publisher.

regression-analysis

The Regression Analysis

What is it? Regression analysis involves analyzing the impact of historical ads on conversion behavior, isolating the contribution of ads in different historical lookback windows.

What does it show you? Conclusions help define strong relationships between conversions and clicks, and conversions and impression for periods of time prior to conversions.

Note: Can require specialized resources to perform this analysis.

 

Conclusion

Setting overly short lookback windows, many advertisers risk missing out on conversions or skewing their attribution in favor of clicks and other lower-funnel tactics.

So how should you address your lookback windows? We recommend starting with a 30-day window for impressions and clicks. Based on our learnings, we believe that this is the proper starting point for many advertisers. For higher-priced, highly considered purchases like automobiles, vacation holidays, etc., a longer window may be more appropriate. For quick, impulse purchases like many eCommerce transactions, a shorter window may be justified. The idea is to capture, analyze and understand the consumers’ purchase journey correctly, not just the last few moments of it.

Over time, using data-driven insights coupled with your understanding of your customer, products and purchase behavior, you’ll be able to adjust your lookback windows to match how your customers consider and buy your products, and ensure you’re making the right decisions with your media budget.