Is All My Direct Traffic Really Direct? Does Source/Medium Accurately Reflect How Users Perceive Their Customer Journey?
Typically, in Google Analytics, to determine where your site traffic is coming from we rely on the source and medium. For example, if a user came from a Facebook ad, the url used in the ad would contain a source and medium (Facebook/ppc). The source and medium would then show up in Google Analytics alongside the purchase, helping us determine where the purchase originated from.
Source and medium give us a good idea of where the majority of traffic comes from, but often, we'll see traffic from a source/medium of (direct)/none, this signifies that the site was visited directly, without a link click. In at least some of these cases it's highly likely that a user clicked on an add, found your site, and then made the purchase on an alternate browser or device that lost its link to the original source/medium. The consequence is we lose attribution for these purchases and they show as (direct)/none. This lost data can have a major impact on quantitative measures like return on ad spend, artificially reducing it, and potentially making campaigns appear unprofitable, when, in fact, they are.
Fairing (formerly Enquire Labs) asks survey questions after purchase to collect this information directly form the user, and get around the problem of lost or missing attribution. With Elevar's data layer, and prebuilt Fairing tags, we can pull this information into GA4, make a report to analyze it, and get some very interesting results.
- Install the Fairing prebuilt container from the Elevar app. Instructions on downloading the prebuilt container and publishing it in Google Tag Manager can be found here. Make sure that you've filled in your Google Analytics 4 ID in the GA4 tag. If you've already published the imported tags, make sure you hit publish again after adding your GA4.
- Ensure the Google Analytics integration is turned on in the Fairing admin settings.
- Your data should now be flowing into Google Analytics 4.
Now that we're collecting data in Google Analytics 4, let's build a custom report that compares the source and medium in Google Analytics 4 to the source provided directly by the customer in the Fairing survey.
- Create a new exploration report in Google Analytics 4
- Select 'Free Form' for technique and add first user source, first user medium and fairing response as dimensions. Then add event count as a metric.
- Drag first user source, first user medium and fairing response into the rows section of the report. Drag event count into the values section of the report.
- The last thing we'll do is filter the events to only include our fairing survey results. Add the event name as a dimension and then click the filters button.
- We'll filter for only the fairing_survey event, choose event name, select match type exactly matches, then write fairing_survey in the enter expression box. Click apply.
- Add another dimension to your report, this time the fairing question dimension. You'll now want to add another filter to ensure we are only seeing results for the question "Where did you hear about us", then click apply. If you changed the question ensure that you are using the exact wording you are using in the Fairing admin.
- To see all the rows of this report click the show rows and select 100. This should be enough to show all results, if not select a higher number.
- You now have a report that compares Google Analytics 4 source and medium to the user declared source. The most interesting rows are those with a first user source of direct. For example, in this report we see 18 purchases that Google Analytics 4 would have attributed to direct traffic, but according to the customers, they first heard about the store via Instagram & Facebook. These sales are likely not attributed in the ad platform (Facebook Ads Manager). With this information in hand a potential actionable item is should we be spending more on Facebook ads?
- Another interesting part of this report is where the fairing response doesn't match with the first user source. For example, row 5 and 6 show that Google Analytics 4 attributes about 13 sales to google/cpc when the customer states that Facebook/Instagram was the real source. With the information from this report we have over 30 purchases with Facebook as the source, that were attributed incorrectly to other platforms, or to direct traffic. It's this information that's so critical in decision making, especially as volumes rise.
This report can be modified to include the answers to other customized questions you've added to your Fairing admin. It's important to note that while this report provides interesting insights we are comparing first touch attribution (first user source/medium) with what the customer considers the initial source they heard about the site from. This is not the last click attribution that we are accustomed to in Universal Analytics.
Updated about 1 month ago