List the blind spots in attribution and the other evidence that fills them.
Open your GA4 acquisition report and note how many recent enquiries or sales are labelled direct or organic search. Then think about the last few customers you spoke to and whether you know how they first heard of you. Attribution misses whatever sits between those two pictures.
Attribution only works on what can be tracked, and in practice that means clicks on links. A large part of how people choose a business never involves a trackable click. Any of these can start or finish a buying decision without leaving a click behind: a friend's recommendation over lunch, a message forwarded into a parents' WhatsApp group, a podcast mention, a flyer at the void deck, a conversation at a wedding, or a shopfront someone walks past every day.
When that person later visits your website, they type your name or search for it, and GA4 records the visit as direct or organic search. The model then gives credit to whichever trackable touch came last, which is usually the one that did the least persuading. Every attribution model from lesson 5.1 shares out credit only among the touches it can see, so the most important touch can be invisible to all of them.
Priya runs a tuition and enrichment centre in Singapore. Her GA4 reports said most new enquiries came from organic search and direct visits, and her Google Ads and Meta ads together claimed a few more. At a trial class she asked a mother how she had heard about the centre. The mother said a friend's son attends, and the friend had sent the centre's number to their class WhatsApp group. Priya then asked the next 5 parents, and 4 of them (example figures) gave a version of the same answer. None of these recommendations showed in her reports. They appeared as direct visits or as searches for the centre's name, and Google got the credit.
The gap is large for businesses that run on reputation. In Singapore this includes most tuition centres, clinics, renovation firms, home bakers and financial advisers. For these businesses, the channels attribution cannot see are often the biggest ones.
Some touches that could be tracked still drop out of the record. Visitors who decline analytics cookies may not be recorded at all, or only as modelled estimates, as lesson 4.2 explains. Their visits never join a path, so they cannot get credit.
Switching devices also breaks the chain. Say a parent clicks Priya's Meta ad on her phone at lunch and books a trial on the family laptop that evening. GA4 treats these as two different users. The ad click sits on a path with no result. The booking sits on a path that began with a direct visit, and the ad gets no credit.
Every model, data-driven included, works with a partial record. It can still tell you useful things about the touches it sees, as long as you remember it does not see everything.
The cheapest fix is to ask. A self-reported source is a question such as "How did you hear about us?" on an enquiry form, booking form or checkout page, and it captures what tracking misses. People are surprisingly willing to answer, especially when it is one short question with a dropdown. A few rules make the answers usable:
Offer a dropdown with a fixed list of options, so the answers can be counted. Write the options in your customers' own words, such as "a friend or family member" or "WhatsApp group", and leave out internal channel names. Keep the list short enough to read at a glance. Add an "other" option with a short text field. Put the question near the end, after the important fields, so it does not put anyone off finishing the form.
Self-reported answers have a weakness of their own. People remember the last or most memorable touch, and they rarely list every touch. Someone who saw 3 of your ads may still say "Google", because that is where they typed your name.
Treat self-reporting as a second view that sits beside your tracking data and read the two together. Where they agree, you can be fairly confident. Where they disagree, the disagreement is itself the finding. Priya's referrals showing up as organic search are an example of this.
Attribution and holdout tests answer different questions. Attribution asks which touches were present before a sale. A holdout test asks whether a channel causes sales that would not have happened anyway. You pause one channel for a set period, or in one area if you serve several, and keep everything else the same. Then you compare results from your source of truth with what you would have expected.
Holdout tests cost money, because you may lose sales while the channel is off. Save them for channels where you spend real money and the attribution numbers leave you unsure. Pick a period that is normally steady, avoiding school holidays, exam season and any competitor promotions you know about. Write a plan before you start, and decide in advance how big a drop counts as a real effect.
Priya could pause her Meta ads for 4 weeks and watch trial bookings in her spreadsheet, which is her own record of results. She would choose a normally steady stretch, keep her other channels unchanged, and compare the bookings with the same weeks last year and with the weeks just before the pause.
If enquiries barely change while the channel is off, the channel was mostly taking credit for customers who would have come anyway. If enquiries drop clearly, it was adding something. Other things can change during the test, such as school holidays, exam season or a competitor's promotion, so read the result with those in mind. Bigger businesses run holdouts more formally, using matched regions or matched groups of customers. For a small business, a careful pause with a written plan is a reasonable start.
For most small businesses, the self-reported question is the place to start. It costs nothing, it collects evidence from the day you add it, and module 8 builds it into unit economics as the second view of your channels.
Your task now is to decide where your question goes and what options it offers. Pick the one form most new customers fill in, whether that is an enquiry form, a booking form or your checkout page. Then write a short dropdown list in the words your customers use, add "other" with a text field, and place the question after the important fields.
Add a how-did-you-hear-about-us question to one form or to your checkout, and write down the answer options you will offer.
Junxiong-WFG Organisation is an authorised representative of AIA Financial Advisers Private Limited (Reg. No. 201715016G).