When I talk to colleagues from online marketing about tracking and attribution, we quickly come across ad blockers, GDPR, cookies and the associated increased requirements. If you go a little deeper into the discussion and look at the traffic channels, it turns out that for many, the channels SEO and Direct are the most successful – i.e. most conversions result – while the paid channels (e.g. SEA and social) stagnate or even go back. The great success of the direct channel is then explained by their strong brand.
We at everysize can also confirm this statement based on our Google Analytics numbers. But why is that? Out of curiosity, I went into the matter and took a closer look at the reasons for our high proportion of conversions in the direct channel. I came across some very interesting aspects, including: The more we scale our paid channels, the higher our traffic and especially our conversion numbers in the direct channel increase!
Attribution != Traffic
Before I get in little deeper, I would like to briefly define what I mean by attribution. Attribution is about assigning conversions (e.g. sales) to an advertising channel so that the success of the corresponding channel can be assessed, especially in comparison to the other channels (see also Exatag: Marketing Attribution Part 1). So it is NOT about traffic (sessions, users
etc.) of a channel, but about the conversions that result in leads and sales and the corresponding assignment. Defnition is important because the entire line of articles always refers to attribution.
“Direct” – The pool of lost users
Everything that cannot be clearly assigned to another channel ends up in the direct channel. It is therefore a collecting basin for all possible channels and is difficult to re-assign to the actual channels. But: the “fuller” the pool, the bigger the problems and challenges.
Of course, I knew of many causes for direct traffic, e.g.
- Type-in traffic
- Links in emails (e.g. Outlook)
- Others, see:
„Direct Traffic in Google Analytics entschlüsseln“ (DE)
„Dein Direct Traffic ist nicht, was er zu sein scheint“ (DE)
„In-Depth Guide on Direct Traffic in Google Analytics“ (EN)
„What is dark traffic in Google Analytics?“ (EN)
However, this does not provide an explanation for the connection with the paid channels and why Direct has so many conversions. In almost all sources that I could find, it was always recommended to use UTM values wherever possible to keep the direct portion low. Since we are already doing this, it could not be the cause.
Why “Direct” is so dangerous for the attribution
The best way to show the danger of a high Direct component is to show an example:
A user clicks on a shop’s Facebook ad. He will be forwarded to the shop including UTM parameters. He doesn’t buy directly, but waits a few hours. After that he add the product to the shopping cart and completes the purchase.
Tracking process without disruption
Google Analytics records the user’s impression and correctly assigns it to the Facebook channel. The user receives a cookie from GA with the id 111.222. A few hours later, the customer adds the product to the shopping cart – GA identifies the customer using the cookie and records it correctly. The customer completes the order process – the dealer sends an order event to GA (e.g. via eCommerce tracking), which is assigned to the customer based on the detected cookie. This correctly assigns the sale to the “Facebook” channel.
Tracking process with disruption
Google Analytics records the user’s impression and correctly assigns it to the Facebook channel. The user receives a cookie from GA with the id 111.222. However, there is a disruption in which the cookie gets lost (causes for this see below). A few hours later, the customer adds the product in the shopping cart – GA cannot identify the customer because of the lost cookie, which is why a new cookie with the id 333.444 is generated. There is no longer a “Source” because it is an internal link – so the impression is probably assigned to the channel “Direct” (or possibly “Other”). The customer completes the order process – the dealer sends an order event to GA (e.g. via eCommerce tracking), which is assigned to the customer based on the (new) cookie detected. Result: The Facebook channel is weakened because the visit is recorded but no conversion is assigned.
Result: “Direct” becomes more and more profitable at the expense of all other channels the more frequently a disruption occurs in the customer journey.
The un-/well-known disruptions
In the course of my in-depth research, I came across various causes that have an impact on online attribution. In this series of articles I would like to go into some lesser known attribution problems and show why many conversions incorrectly end up in “Direct”:
- Privacy blocker – the unknown danger (Part 2)
- Does the “private mode” eat the cookies? (Part 3)
- The Reason for increasingly poor conversion of paid channels on mobile (Part 4)
- How Google Tag-Manager and tracking-swiches distort the attribution (Part 5)
- Why cookie consent layers are not a solution (Part 6)
- Hope dies last – my attempt to conclude (Part 7)
The challenges are increasing
To anticipate one thing – I’m not a pessimist and I do NOT want to initiate the downfall of online attribution with this series of articles. However, I would like to clarify that the challenges for proper tracking are increasing. New technical solutions are necessary and the perspective on the assessment of the marketing channels must also be reconsidered. In the future, simply integrating Google Analytics or Google Tag Manager everywhere will no longer suffice.