Attribution for Small Businesses: Which Channel Worked?
By CodexierPublished 5 min read
Every small business that spends money on marketing eventually asks which channel actually brought in the customers. Analytics tools offer an answer, ad platforms offer another, and they rarely agree. That is not a bug in your setup; it is the nature of attribution. Customers see an ad, search later, hear about you from a friend and call from the Google Business Profile, and no tool sees all of it. This guide explains the models, their blind spots, a low-tech method that fills the gaps, and how to decide with imperfect data.
Why attribution is hard
Attribution breaks for three reasons. Journeys are long and cross devices, so the phone that saw the ad is not the laptop that filled in the form. Consent rules mean that visitors who decline cookies are only partly measured, which is correct under Swedish law but leaves gaps. And each ad platform counts conversions its own way, crediting itself generously, so the totals across platforms add up to more customers than you actually got.
Last click, data-driven and others
| Model | How it credits | Strength | Weakness |
|---|---|---|---|
| Last click | All credit to the last channel before the conversion | Simple, stable, easy to explain | Ignores everything that created the interest |
| Data-driven (GA4 default) | Spreads credit using patterns in your data | Accounts for several touches | A black box; needs volume to be meaningful |
| Ad platform attribution | Credits its own ads, including views | Useful for optimising within the platform | Overlaps with other platforms and overstates |
| First touch | All credit to the first channel | Shows what brings new people in | Not available as a model in GA4; must be approximated |
Google retired several older rule-based models in GA4, such as linear and time decay, leaving data-driven and last click as the main options.
For a small business with modest traffic, the model matters less than consistency. Pick one view for your monthly report, stick with it, and interpret changes over time. Our GA4 setup guide covers how to configure conversions so the numbers are at least measuring the right things.
Blind spots: phone, word of mouth, dark social
Phone calls
Calls from your Google Business Profile or a number on the site are not website conversions unless you track them. Call tracking or a separate number per channel can help.
Word of mouth
A recommendation from a friend shows up in analytics as direct traffic or a branded search, and the channel gets the credit that the friend earned.
Dark social
Links shared in chats, Messenger, WhatsApp or email arrive without referrer information and are counted as direct.
Declined consent
Visitors who decline analytics cookies are measured only in aggregate, if at all. Consent Mode can model some of the gap.
The gap from declined consent is handled partly by Google's modelling, which our guide to Consent Mode v2 explains. The other blind spots need a different method.
Asking customers directly
The simplest attribution tool is a question. Add an optional field to your contact form, booking flow or checkout: how did you hear about us? Use a short list of options plus free text, and store the answer with the lead or order in your CRM. For phone enquiries, have whoever answers ask the same question and log it. Self-reported answers are imperfect, since people remember the most recent or most memorable touch, but they capture exactly what tracking misses.
- Offer five to seven options: search, Google Maps, social media, recommendation, an event, a partner, other.
- Keep the field optional so it does not reduce form completions.
- Save the answer in the same record as the lead, so you can see revenue per answer.
- Review the answers monthly next to your analytics report and look for disagreements.
Making decisions with imperfect data
- Track leads and revenue per month in total, which is the number that cannot lie.
- Watch each channel's trend in one consistent model, not the absolute split.
- Compare with self-reported sources to spot channels the tools undercount.
- When in doubt about a paid channel, pause it for a few weeks and watch the total.
- Decide on the direction the evidence points to, and revisit next quarter.
When you do not need help: with a single paid channel and a simple site, the steps above can be done in an afternoon. Help pays off when you run several paid channels, have phone-driven sales or the numbers from different tools disagree wildly. Our tracking and analytics setup configures conversions, consent and CRM source fields; see pricing or book a call.
Frequently asked questions
Why does Google Ads report more conversions than GA4?
Ads counts conversions with its own rules, including view-through conversions and different time windows, and credits its own ads. GA4 shares credit across channels. Both can be right by their own definitions, which is why you should not add them up.
Which attribution model should a small business use?
The GA4 default is fine for most. What matters more is using the same model consistently and complementing it with a direct question to customers about how they found you.
Is it worth buying an attribution tool?
Rarely for a small business. Dedicated tools need volume and several paid channels to add value. A well-configured analytics setup, a CRM source field and occasional channel pauses give most of the insight.
Does asking how customers heard about us conflict with GDPR?
No, as long as it is optional, stored with the lead for a stated purpose such as marketing evaluation, and covered in your privacy notice. It is often less intrusive than tracking.
Find out which marketing actually works
Tell us which channels you spend on and how leads reach you today. In fifteen minutes we can point out the biggest blind spots and how to close them.
Book a free 15-minute call