Marketing attribution for small business: consistent beats accurate

Aug 18
3 min read

Updated: 4 days ago

Introduction

Attribution is the question of which channel deserves credit for a customer. It has no clean answer, because customers encounter you several times before buying and no system sees all of it.

That is not a reason to give up on it. It is a reason to aim for consistency rather than accuracy, which is a much more achievable goal and nearly as useful.

1. Marketing attribution for small business: pick one rule

The main options are simpler than the literature suggests.

First touch credits whatever introduced the customer. Good for understanding what creates demand.

Last touch credits whatever preceded the purchase. Good for understanding what closes.

Ask them credits whatever the customer says. Imperfect memory, and the only method that sees word of mouth.

Choose one primary rule based on which question matters more to you now. Then hold it steady — changing the rule mid-year looks exactly like a change in performance, and you will misread your own data.

2. Consistency is what makes it useful

You are not trying to discover the objective truth about each sale. You are trying to tell whether this month is better than last.

For that, a consistently applied imperfect rule works. An accurate method applied inconsistently does not. This is the single most important thing to understand about attribution at small scale.

Write your rule down, including the time window you count, so whoever produces the numbers next month does it the same way.

3. Ask customers, because it is free and it sees what tools cannot

Add one question at enquiry or purchase: how did you hear about us? Record the answer in the same list as everything else.

People misremember, and someone who saw an advert then searched your name will often say "Google". Over months, though, the pattern is genuinely informative — and it is the only way to see referrals and offline mentions, which for many small businesses are the largest channels.

4. Give each channel its own destination

Much of the ambiguity disappears if the source is unmistakable at the point of arrival.

A distinct landing page per campaign, a separate phone number, a code to mention, a dedicated form. Each one removes guesswork rather than resolving it afterwards.

Call tracking numbers are especially worthwhile where the phone is the main route, because that is usually the biggest blind spot.

5. Tag links consistently, or not at all

For online traffic, campaign tags let analytics separate sources. The value depends entirely on consistency.

Decide a naming convention — channel, campaign, creative — write it down, and use it every time. Inconsistent tagging produces a report full of near-duplicates that cannot be totalled, which is worse than no tagging because it looks like data.

6. Test untrackable channels by pausing them

Sponsorship, print, local visibility and word of mouth resist attribution entirely.

Rather than assuming they contribute, test by absence: pause one for a defined period and watch total enquiries. If nothing changes, you have your answer. If demand drops, you have measured the contribution more honestly than any tool could.

Uncomfortable, and the only reliable method available.

7. Do not over-invest in precision

There is a point beyond which attribution work costs more than the decisions it improves.

For most small businesses that point arrives quickly. Knowing that one channel produces customers at half the cost of another is enough to act on. Knowing the exact split of credit across four touchpoints changes nothing you would do.

Spend the effort on recording sources completely rather than on modelling them precisely.

Conclusion

Choose one attribution rule, write it down, and apply it the same way every month — consistency matters more than accuracy at small scale.

Ask customers directly and record it, give each channel a distinct destination, tag links to a written convention, and test untrackable channels by pausing them. Then stop: precision beyond the point of changing a decision is a cost, not an improvement.