How did you hear about us?
Supplement your attribution data
Hey there, it’s Jacob at Retention.Blog 👋
I got tired of reading high-level strategy articles, so I started writing actionable advice I would want to read.
Every week I share practical learnings you can apply to your business.
Attribution is hard.
Where are you users actually coming from?
Is it as easy as just asking them?
Yes and no.
It’s become more and more common to ask during onboarding "How did you hear about us?” (HDYHAU)
Can this replace your other attribution methods?
It’s an important piece of the puzzle, but doesn’t solve all your problems.
I pulled in some insights from some user acquisition experts to get a set of diverse opinions:
Sylvain Gauchet is a Full-funnel growth advisor and consultant for subscription apps and writes one of the best newsletters in the business, 💎 Growth Gems. (He’s hard at work mining gems everyday, that’s why he looks like the way he does…)
Thomas Petit is one of the most knowledgable people in the mobile growth space. And he is very generous with sharing his insights. Follow him on LinkedIn or X to get a little smarter (and at a minimum, be entertained).
Jakub Chour is expert in all things mobile and B2C growth. He’s done everything from CRM to UA to ASO to product growth and conversion optimization. Connect with him on LinkedIn here (He also has a great name 🙂 ).
So why do we need a “How did you hear about us?” question?
Thomas:
Attribution is by nature imperfect, even more so in these post-ATT times. The IDFA-based measurement was never perfect, but the simplicity to rely on it made it the standard for most.
Any last-click based attribution (the most common) is flawed. There is no perfect substitute, and I recommend to embrace complexity by comparing different methods to make more informed decisions.
Those methods may include:
• Network self-reported metrics (usually based on SKAN, fingerprinting, IDFA-consenting cohorts, modeling or most often a combination of the 4)
• SKAN
• App Store Connect traffic sources (with or without CPPs)
• Ideally, using on top an incrementality tool such as Incrmntal
• Or in some rare case MMM (media mix modelling)
Each has their own limitations, but having more information enables better decisions on budget allocation. (For campaign management, it's hard to go beyond self reported metrics, sadly.)
HDYHAU is a complementary method on top of the above. In my opinion, never a real substitute, given user’s perception may differ from the advertising spend, and some channels are harder to match (in particular gaming adnetworks, DSP, Admob, etc.)
Beware that the design & wording of the HDYHAU screen (or screens plural, some may ask further, like which specific keyword for search, which placement for FB/IG, etc) can have a significant impact on the results.
Thomas shared some additional insights in last week’s newsletter here
Where should you put your HDYHAU questions?
I usually recommend closer to the end of your onboarding flow because it’ll have less of a negative impact.
It’s not really adding value to your users experience, so if it’s later in the flow, users who remain aren’t likely to be phased by a question of that type.
Sylvain:
Having it early in the funnel doesn’t do anything for UA, it’s more of an onboarding optimization topic.
Several apps like Duolingo have it right away, others right before email sign up. You have to test what works best.
Let’s look at where a few other apps place these questions.
Babbel has it on their 12th screen:
Jakub:
What I think is often missed with HDYHAU surveys is the overestimation of people's memory (e.g., whether they will accurately recall the right source) or their willingness to answer thoughtfully (people will click anything to avoid exerting mental effort).
That's why it's useful to include a "lie" option, such as "TV."
Additionally, including a choice like "I used the app before" is a smart approach, as it helps filter out users who might otherwise select a source they either don't clearly remember or that is no longer relevant to your current efforts.
RISE has it on their 9th screen:
Sylvain:
Age is good to have along HDYH to compare with what you see in Meta.
The most useful I’ve seen age be is when launching new channels (or campaigns with new optimizations or audiences) in order to understand how the traffic differs compared to what you usually have. Typical example is launching a new channel and conversions are shit: strong chances are you’re bringing in younger users than usual.
Oh and Age can also be used to create a more restrictive optimization event. Example: signups > 18, trials > 18
One reason to have it very early would be if you optimize for something like “age + specific answer” and want maximum reach.
Paired has it on their 12th screen:
Are you a competitor to Paired? My takeaway would be to try to promote my app to therapists and counselors. It’s a meaningful enough channel for them add it.
Sylvain:
I like comparing the evolution of HDYH answers over time to spot anything that stands out and get a read on launching new channels (e.g., TikTok) or initiatives (e.g., influencers). I’ve also seen it used as input to a formula to attribute TV/radio efforts (otherwise it often goes unattributed i.e., organic, or SEM brand).
Myfitnesspal has it on their 11th screen:
Are you running a campaign with Kate Hudson? Make sure you add it to your HYDHAU screen.
This brings up an important point - can you update this screen remotely? Are you stuck to only adding options with an app update? Remote config could be useful here based on the velocity and scale of your marketing.
Thomas:
Interestingly, this screen can even serve another purpose. Beyond gauging attribution, the response can drive significant shifts in monetization, activation, and retention patterns across user segments.
If so, this presents an opportunity to personalize the onboarding experience, paywall, and home screen based on the user's answer, optimizing engagement and conversion rather than using a local maximum.
Endel has it on their 6th screen:
Here are some of my takeaways:
You can research competitor’s HDYHAU to understand channels they’re running
Most of them have TV as an option. Maybe they’re doing more advertising via OTT ad platforms, but I’d bet most aren’t doing much TV advertising. Measure these responses vs. what your attribution tool says.
Hopefully, this is obvious but make sure to randomize your question order! If you don’t all your data will be skewed by the order.
Attribution is difficult in this modern day, so the more data sources often the better to try to get a more complete picture
Age data can also be very valuable in complement to your attribution source data
Test where you should put your HDYHAU us screen to optimization your conversion funnel. There is no one size fits all answer
And one last parting thought from Thomas that emphasizes attention to detail is important:
Don’t forget that wording is critical here. The microcopy can change the answer distribution A LOT. Some options are simple (TV?), some are not at all.
For search & the appstore in general: do you mark "appstore"? "search"? Does search include brand or not? And cover appstore & google web?
Do you need to know featurings vs search? ASA vs not?
Hard choices to make!!
Thank you again to Sylvain, Thomas, and Jakub!
Follow Sylvain on LinkedIn and subscribe to Growth Gems
Connect with Jakub on LinkedIn
Looking to go to a conference this year?
Mobile Apps Unlocked is still the largest mobile app focused conference in the business.
I was a speaker last year, and will be a media partner this year with Retention.Blog - how cool! I would love to meet some of you there!
And most of you will be eligible for free passes if you sign up before April 10.
MAU is doing some new cool stuff this year like their Clubhouse. Want a discount? I secured some promo codes for you:
RETENTIONCLUBFD (25% full day passes; value $75)
RETENTIONCLUBHD (25% half day passes; value $50)
RETENTIONAGENCY ($300 agency passes)
RETENTIONVENDOR ($500 solution provider passes)
Here’s some info on the Clubhouse: MAU Clubhouse | MAU Vegas 2025
📣 Want to help support and spread the word?
Go to my LinkedIn here and like, comment, or share my posts.
OR
Share this newsletter by clicking here.
