There may not be any other feature in Google Analytics that I feel more strongly about promoting than this. A topic that I have discussed a lot during my eight years of measuring marketing through Google Analytics as both a agency professional and web analyst at a corporate advertiser.
The lack of knowledge agitates me as it is very beneficial for Google, but on the expense of your business. Resolving this issue through changing your setting will move you from the current state where Google is overattributing their own ad products, and towards a future state where your analytics are a greater representation of your actual situation. Let's get on and solve this problem once and for all.
What you will learn in this article:
- What is campaign timeout in Google Analytics
- Why this crucial feature is in need of your attention
- How to fix it
- What others business think and do
- What is the best practice, adapted for your business
What is campaign timeout in Google AnalyticsCampaign timeout is a feature in Google Analytics where you can define how many days it can pass from a user clicked a campaign ad, and a later conversion still gets attributed to the campaign. All marketeers today acknowledge the consumer buying process, and that we work with consumers in differents stages of this process. Even though we want everyone to convert immediately, we often just get a click and a visit. If the users comes back and convert at a later point, we would still like to give credit to that campaign. Campaign timeout is where you define an acceptable time window.
How do other businesses decide the length of their campaign timeout setting?Most Analytics accounts I have seen has this on default settings. In an untouched state, most likely because they don't know about it's impact or have just ignored it for now. This is horrendous as you later on will take business critical decisions based on your reports with numeric results calculated on this basis. And having this set to 6 months is clearly beneficial for Google, but not necessary remotely correct for your business.
Those who do adjust this setting usually goes for 30 days which has been the standard in the web analytics world. Experts explain this length as still being a quite randomly selected number, just because a month is a easyily communicated time frame. Read on to understand why this is not recommended.
How does this work in other tools than Google Analytics?Each tool that displays you comversions in relations to it's own ad delivery, will have it's own definition of how to attribute conversion. This is one of the major reason behind the confusion around why all these tools show different conversion data.
- Facebook Ads is able to give you overview of a web page conversion related to your campaigns. The default setting here is 30 days on click and 1 day on impression. The latter here is attribution only based on someone seeing your ad, without clicking. You can easily toggle both of these settings in the ad administration tool, but be aware that it may drop back again to default setting next time around.
- Google Ads also gives you overview of your conversions in relation to your ads. If you have linked up your account with Google Analytics, then you will be able show the identical goals in both tools, which is very nice feature. Keep in mind that the length of the conversion window is set independently per eachs of these tools. Check your setting.
Another thing businesses do wrong when comparing campaignsFor a fair comparison both campaigns should have an equal long longtail, where the timeout period has run it's course. E.g. you should not yet compare the summer campaign which ended yesterday with last years campaign. Last year campaign will then have the full longtail effect as a result of your timeout setting. Make sure you compare campaigns when both have ended (no more clicks) and when the campaign timeout period has expired for both!
How should I decide the length of my campaign timeout setting?
- Start playing around with a number
Starting on, let's say 4 weeks, will get the discussion started. Does it initially feel too long? Too short? Why does it feel this way?
- Think in weeks - not in days or months
To make sure you count the same amount on unique weekdays.
Consumers are humans. Most of us humans live our lifes on a somewhat weekly basis. Like working during the week and having days off in the weekend. Because of this, most consumers display somewhat different behaviour during midweek versus weekends.
30 days is not a good setting as this will contain four full week cycles + two more random days. If you compare two months to each other, one may have 4 weekends and the other may have 5. This may have great impact as I learned when working in a lottery company where 50% of their weekly revenue was generated on Saturdays. Having an extra Saturday in March would make that a monthly report look way much better than April where these extra days landed upon the midweek.
- Think about your customers and their decision making process
The nature of a consumers relationship with your type of product/service is highly relevant. This is examplified in the following three examples:
- 7 day: A weekly lottery with drawings each Saturday. Part of a weekly routine. Buying cycle is frequent and the decision making process is spontaneous. Cost of investment is low.
- 28 days: Online clothing store. Buying cycle neither very long, nor short. Seasonal changes and time-limited offers may speed up, but time passes easily as the urgency for the consumer is quite low. The decision making is easy to postpone. Consumers spend time in the comparison stage and may be awaiting in case a good offer pops up. Cost of investment is medium.
- 6 months: A car dealership. A lengthy decision making process for consumer. In addition to cost of investment being high, the consumer's is deciding upon a long term relationship with the product, hence a greater importance. Decision-making-process is stage-based rather than impulsive and spontaneous.
- This one setting will cover all your products and services
You may offer different products that has different characteristics regarding it's consumer buying process. Unfortunately, this setting can't be differentiated amongst your products so let this be part of your consideration.
- Let your current marketing sitation affect your decision
If your business is hidden away and only found when running digital campaigns, then a longer time period would be preferable. The only way a consumer would revisit your page directly would be because of previous touch points with your ads. No risk attributing to wrong touch point. In an opposite case, where your company has a lot of brand awareness, this has been built up through many touchpoints over long time. To attribute 100% of a later conversion to one single click on a recent Facebook Ad is just absurd. In this cases it's preferable with a quite short timeout period to avoid false attribution.
If your marketing mix contains traditional marketing, e.g. TV/Radio/boards, you know that these activities are way harder to track in regards to conversions. The lenghtier you set this time period, the more likely it is that your digital campaigns takes all the honor even though it was your recent TV ad that stimulated to that spike in webpage visits.
- Discuss with your colleagues
This is highly recommendable as it will spike new thoughts and considerations. It will also challenge you to more clearly define why you think the way you do. Another great value of this is that you hopefully reach a mutual agreement upon this foundational rule upon which you will base all future campaign evaluations.
Think in amout of full weeks rather than days or months to keep the count of unique weekdays fair. Your alternatives should therefore be: 7, 14, 21, 28... and so on.
What will happen to my numbers if I change this setting?Changing this setting will have a big impact to the numbers, but only in the campaign reports as this is only how conversions are attributed to campaigns. The totals on other non-campaign reports will stay the same. Most importantly we here try to define a timeperiod rule, which in turn will set all campaigns on the same level. This level will be what we gradually get used to when comparing our future campaigns. The best advice is to get clear on what objectively would be the best setting for your business.
It's rather quite strange how many users in Google Analytics tend to keep settings as they are, even though they know it is wrong. Maybe this has something to do with a general reluctance to change, and uncertainty to what new numbers may say.
Personally, I'd rather get my Google Analytics straight starting from NOW. Historic data and previous reporting levels are not much worth for future comparison if they were all wrong in the first place.
Changing this setting will NOT affect channel grouping ( "social", "direct" etc) as numbers here are session data, based on how they actually arrived to your site during that specific session.
Where do I actually do the configuration changes?Simply open Google Analytics and go to Admin. The center pane is for Property settings. Here you will find Tracking info which contains Session settings. Don't get fooled by the title, as you can change both Session and Campaign timeout settings in here. Adjust your settings carefully and press Save.
Tip: It's best practice to keep filtered main reporting view as well as a second view with all default GA settings and no filters. Just as a backup with all data intact. In addition to these two views, a view for testing purposes are often recommended. Keep in mind that Session and Campaign timeout settings is on a property level and will affect all your different Google Analytics views.
SummaryCampaign timeout setting is an important in Google Analytics that is widely ignored. An alarming truth, since this setting is an essential part in the calculation of how much conversion was created by your digital marketing campaigns. Keeping this set to it's default setting (6 months!!!) is most likely horribly wrong for your business. The consequence is that you may perform business critical decisions upon reports which are giving you a false representation if your actual situation!
A too short campaign timeout period will underattribute, but at least it becomes quite accurate with the fewer result that do appear.
A too long campaign timeout period will overattribute, letting your digital ads take the whole honor for a conversion even though many other things can have happened during these months. Leading you into pooring more money into your next digital campaign, though that may not yield the expected results. Amonst the digital ads, paid search ads are probably those getting away with most of misattributed conversion. Google Analytics is attributing based on last non-direct touchpoint, and in a multi touch point scenario an intentional product search is more likely to be occure last before the actual conversion takes place.
Share this article with your colleagues, sit down and have a heathly discussion. Multiple brains work better than one and a discussion can both spike new thoughts and those old ones might need to get challenged from time to time. When you reach a decision, figure out when you will do the needed configuration change and add an annotation. This brilliant feature in Google Analytics will add an icon on the timeline so everyone can see when this change was done when they in the future look at historic data.
Wish you a happy analytical day!