Choosing the attribution for email, push, in-app, SMS campaigns seems very simple – pick a period of 3-7 days and make a report. Well, as everything, once you dig a little bit deeper, it’s not that simple.

In this post, I will argue that not all campaigns should have the same attribution window, not all messaging channels, and not even the same campaign sent on a different day.

Before I start, I want to make a disclaimer that my explanations will be based on ad hoc campaigns for simplicity. The same principles apply to lifecycle campaigns, but there is an additional dynamic as each email is sent at a unique time.

What is an Attribution Window?

In marketing, attribution window is a term used to describe a period used to measure the campaign’s impact. Most frequently, it’s going to be defined as a period of 1–7 days, but in some specific cases it can be a period of a few hours, or even weeks.

Let’s say we want to send an email on Monday and measure the effect it had on Wednesday. We just defined a 3-day attribution window for that email, as illustrated by the image below.

Illustration of a 3-day attribution window for an email campaign

Email data will continue to be generated on days 4,5 and 6, but in this case we opted-in to disregard it as we think the bulk of the campaign’s effect happens in the first three days.

Why Make a Fuss About Picking the Attribution Window?

You might be thinking that all of this is unnecessary. Why does the attribution period matter? The campaign’s effect should be visible no matter when we measure it, a day, two or five after it’s sent, right?

Not right. There are a couple of very compelling reasons why an attribution window should be carefully chosen:

Miss-attribution. Using too long of a window means it’s likely to overlap with other campaigns, which will pollute the data used for measurement.

Illustration of miss-atributiion due to overlapping attribution windows of two emails

The campaign’s effect is not the same through time. The initial effect is usually the highest, and then fades towards the baseline, or even below it.

Ilustration of conversion rate variablity impacted by an email campaign

    How to Define the Appropriate Attribution Window?

    The right answer is a combination of different methods that intertwine to achieve an optimal solution. Let’s explain each of them.

    Pick a period until the next campaign is sent to the same audience

    Let’s go back to our Monday email, for which we want to measure its impact on purchase conversions. If we have another campaign going out on Thursday that even partially overlaps with the same audience as the Monday one, and can also impact purchase conversions – that defines our maxim attribution window as 4 days. This is because we don’t want to misattribute conversions from a Thursday campaign, to the Monday one.

    Align the attribution period with a real-world event

    If your conversions align with a time-sensitive event in a real world, it makes sense to take that into consideration when choosing the attribution window.

    Let’s say you’re sending a push notification to promote a basketball game at 9pm. As a typical basketball game lasts around 2.5h, that gives us information that our attribution window should be only last until the game ends – a few hours only. Measuring any longer makes no sense as user can no longer watch the game.

    Attribution window based on what messaging channel are you using

    I think different channels inherently come with different attribution window ranges due to the nature of how they are consumed.

    • Emails land in the inbox and can just sit there. Based on my experience, the most opens happen on the day of the send, but a significant chunk happens on the latter days also. That is why email supports longer attribution windows, usually up to a week.

    • Push notifications tend to be engaged within a few hours of receiving them – the user is either going to open it, or swipe it away. It’s not sitting in an inbox somewhere, so it can be revisited and clicked-through again. That is why I think push notifications allow for a relatively short attribution window – only a few hours, or until the end of a day. When a user wakes up the next day, there is a very little chance that notification has any effect.

    • In-apps pop-up and interrupt user experience. They require immediate interaction with them – either click the CTA, or close it. It can’t even be ignored for a little bit like a push notification can. That is why I recommend short attribution windows for in-apps – consider counting only conversions that happen during that session, within a few hours or up until the end of the day.

    • SMS and Instant Messaging (WhatsApp, Facebook Messenger…) do not require instant interaction, but they have incredibly high open rates – 97% or so. Most of the users will open them the same day, but similarly to email, some will still do it days later. For that reason, it makes sense to give it a bit more time and opt-in for an attribution window of a few days. Ultimately, you can support this decision with data – look at your open rate distribution and cut it off at a point where no, or very few opens are coming in.

    If you want to know more about general advantages and disadvantages of these channels, you can check my previous article Email vs. Push vs. In-app vs. SMS – What to Use and When.

    Attribution window based on the natural conversion interval

    How long does it take for a user to convert from a specific campaign? Let’s say you are promoting something for an ecommerce store, you’re selling small items like clothing and your conversion event is a purchase. For a user, deciding to buy clothes if they see something they like is relatively easy, and you can expect the bulk of conversions to happen within a day or two of sending the campaign – that means we’d use a short attribution window.

    Now, let’s say you also run an ecommerce business, but you’re selling big, expensive items like furniture. In such case, the user will usually take more time to inform themselves about the product, take room measurements, consult their family, wait for their next paycheck, etc. All those actions prolong the buying process, which means we are to use a longer attribution window, maybe even up to two weeks.

    Analyze Historical Data and Trends

    All the previous tips can be amplified by looking at your previous campaigns and seeing what attribution window should you be using for your campaigns. This is very important as every business is unique, so this to personalize it even more. Personally, I don’t think it’s completely necessary, as this takes even more time to implement. If you don’t have the time, logic can get you in the ballpark pretty accurately,

    The downside of using a variable attribution window

    I’ve described a lot of benefits of deciding the attribution window per-campaign basis, but it’s important to mention one important downside that comes with it – compromised benchmarking.

    Normally, we all compare campaigns against each other in order to draw insights and make improvements, but that becomes compromised if each campaign has a different attribution window. Let’s say one email has a window of 1 day, while the other has a window of 7 days – the second one is likely to always have higher conversion rate and engagement just because it had more time to collect data, not because it’s necessarily a more successful campaign.

    The Conclusion

    In the end, my opinion is that picking a 3-day (or any day) attribution window and using it across all campaigns has its place. If your email campaigns are few and simple, you have too much work already, or you value benchmarking too much, I think that picking a number and sticking with it consistently is a good practice.

    On the other hand, if you run a serious and ambitious customer engagement practice, you have bandwidth, and you want to improve your measurement process to get more accurate insights from your campaigns, give my recommendations a try. It doesn’t have to start complicated, try it with a campaign here and there to see how it goes. If you decide to stick with it, deciding the right attribution window will become instinctive and you won’t have to give it much thought.

    Ultimately, the goal is to understand the nuances of your campaigns and make data-driven decisions that align with your objectives. By carefully selecting the appropriate attribution window, you can gain clearer insights into the true impact of your campaigns and optimize your marketing efforts for better results.


    How did you like this post?

    Thanks for your feedback!

    🚀

    Why name denis-test.com?

    © 2024 Denis Kolbas

    Privacy Policy