There are some common questions occurring in marketing forums: why is the source information blank in a user’s account? Is there a UTM tag tracking issue on Google Analytics? Why does increasing ad spend on best-performing social channels not correlate to increased signups from that source?
All these questions have one thing in common: issues with user attribution.
The short answer to all these questions is that more users are blocking tracking cookies, which means that attribution is not connected to a user. You may have seen this when running an ad campaign, it resulted in more sign-ups but there is no campaign attribution connected to the user.
This could be that users are using browsers like Firefox that block tracking cookies by default, or they are actively making the decision to not be tracked by clearing their cookies, using a VPN, or changing their privacy settings.
It means that marketers need a better strategy when it comes to sourcing attribution. Here we’ll break down each step of attributing users who have blocked tracking.
Stage 1 – Checking your attribution set up
In this first stage, you’ll need to do a little detective work and look at what attribution you already have set up.
- Are conversions registering?
Many tools including Google Analytics don’t allow for historical data to be pulled from a conversion – the conversion only tracks once you’ve got it set up. In this case, conversions should still be coming through but they may be registering as ‘Referral’ or ‘Direct’ traffic.
If you can see users signing up but there are no conversions registering at all, you’ll need to fix this before going any further. It is likely an internal issue.
- Is there a trend to the attribution drop off?
You’ll want to check if there is drop-off of a particular audience segment. For example, Apple rolled out the change in iOS 14.5 where an alert is issued when an app is downloaded, asking the user to allow said app to track their activity. In the first three weeks, there was a consistent trend of around 70% of worldwide users not opting into app tracking that is used for marketing purposes. If you’ve seen a considerable drop off in user behaviour on an iPhone app since 26 April 2021 when iOS 14.5 was released, it’s likely that this may have played a part.
If there are any key dates where large drop-offs in tracking started or if you notice that certain popular devices like iPhones are excluded from your reporting, be sure to check changes in data privacy against the points you’ve identified.
- Has there been any issues with other teams in your business?
Product, tech, brand or other departments may have noted gaps or different behaviour from users. It’s important to use cross-team data to pinpoint issues.
It’s what Uber should have done to identify why paid ad spend was not increasing user acquisition. In 2017, Uber was under a lot of scrutiny due to allegations of sexual harassment and unethical practices during travel ban protests and was losing thousands of customers despite spending $150M on ads. Uber’s ads were appearing on far-right online media outlets despite asking their media vendors to not buy ad space on those sites.
In order for Uber to see what media vendors were still running those ads on blocked sites, they paused certain ad groups. They found that there was no change to user acquisition despite pausing millions of dollars of ads. After more investigation, Uber found that their media vendors were misleading clicks and attribution and that most of Uber’s customers were coming in organically from mechanisms like user referrals. Not only did it mean that other teams’ efforts were contributing more than was estimated, but they could also reduce their paid ad spend by $100M.
This issue was only raised to Uber by watchdog Sleeping Giants who continued to flag ads appearing on those far-right sites. This is not the only way that Uber could have identified the issues around media spend. There would have been user behaviours identified across teams for example the strong performance of referral traffic. That’s why it’s important to check across teams before moving into the next steps of checking attribution.
Stage 2 – Deciding what attribution you need
Now you’ve identified how your current tracking works, it’s time to start structuring your attribution strategy. You’ll need to answer the following.
- What are you using attribution for?
Broad or highly specific attribution will depend on how mature your business is. For example, a startup with one person in marketing that’s running a $3K monthly social advertising budget only needs to know what platform is getting the most sign-ups. However, an online news website would need more longtail tracking as they want to know when someone read multiple articles or performed certain actions on their site in order to serve them with more relevant content.
Clearly define what you need tracked and why, as this will be the base of your attribution strategy.
- How much time/effort do you have to spend?
Make no mistake, attribution is important but it can quickly become an area where you spend significant time. This can cause delays to other projects going live and the growth of your audience.
Treat attribution like any campaign with a set live date. This will clarify solutions that meet your timeline or allow you to weigh the time spent and if it’s worth getting external support.
- Is the tracking in the user’s best interest?
Attribution is great at giving more relevant information on how the user discovered your product or being able to show them your product in a relevant context through retargetting. However, you need to ensure that any retargeting pixels or source tracking is done legally.
Your attribution should always have the user’s perspective in mind and that they have all the information or opt-out mechanisms accessible to them.
Stage 3 – Deciding on a solution
From all this information, there are two key solutions.
- Manual solutions including segmenting sources or user’s self serving signup information
If you don’t want to add more tracking technologies, you can use lo-fi solutions. To see what sources are converting the best, once you have a stable number of sign-ups, you can test running more budget towards on ad channels or turning off channels one by one to see if there are any changes to your number of sign-ups. This will give you a broad sense of where unattributed conversions are coming from.
The other option is to ask users when they register how they heard about your brand. Depending on your CRM or sign up process, you can only trigger these communications to send to users where the source is blank.
Manual solutions are good for businesses who want quick solutions that give a broad-level understanding of attribution. It also starts the conversation with the user when they sign up, shifting from passively tracking information to users being more involved with how and why they are receiving communications.
- Technology solutions including tracking session data
Other solutions involve changing how you track data. For most marketers, you’ll be tracking users with cookies but you may want to consider tracking session data at key events along the user journey. Where cookies store data in the user’s browser, session data is kept on the server. The server in this case may be your internal platform, marketing automation platform, or analytics tool.
It means that you’re not as at risk if people clear their cookies or turn off tracking as it’s attributed to their user session. However, session data can mean slower websites due to the engineering needed to change tracking from client-side to server-side, and to maintain this tracking.
For this, you may need to work with your product or tech teams. Their solution will depend on your brief on what you are needing to track and why. If you are using Google Tag Manager Server-Side containers, you are already tracking session data and may only need to update which events you are tracking.
How can I future-proof user data attribution?
Technology is set to change. There are new solutions like Unified Ad ID in which a user has an anonymized identifier and can set their preferences for a supplier network rather than having to opt-in or out every time – however, Unified ID 2.0 is still not live and there is no consensus if this should be the solution.
There is the rise in easier ways for users to stop tracking or to go incognito, so while short-term solutions like changing tracking may help, focusing on how users can give feedback or auditing what things you actually need to track will help you future-proof your work against changing technologies.