The Honest Truth About Marketing Attribution for Small Business

Why perfect marketing attribution for small business is impossible

Start with the plumbing. A visitor reads about you in a newsletter on their phone, forgets, googles your brand a week later on a laptop, clicks nothing you can tag, and buys. Which channel earned that sale? Every honest answer is a guess. Four forces make it that way, and none of them are going away.

• Dark social: the link someone pastes into a WhatsApp group, a Slack, a DM, or reads aloud to a friend on a call. It arrives with no tag and often no referrer, so it lands in your reports as direct or as nothing at all.
• Direct traffic: a growing share of genuinely referred visits get filed as direct because browsers, apps, and AI assistants drop the referrer on the way in. A fat direct bucket is not loyalty. It is missing data wearing a disguise.
• Multi-touch journeys: one purchase might touch a podcast, a search, a retargeting ad, and an email over three weeks. Any model that hands all the credit to one of them is picking a story, not measuring one.
• Privacy loss: tracking prevention, cookie limits, and consent choices mean a real slice of your traffic is uncounted or stitched together from assumptions. The trend runs toward less visibility every year, not more.

I learned this the slow way. A launch I was certain came from Twitter showed up almost entirely as direct, and it took me a month to accept that organic and referred visits routinely hide in the direct bucket. The channel I could see was not the channel doing the work.

So should I stop tracking attribution at all?

No. Giving up is the other failure, and it costs just as much. There is a wide gap between perfect and useless, and almost all the value lives in the middle. Directional truth beats decimal-point fiction. I do not need to know a channel drove 34.2 percent of revenue. I need to know whether it is clearly working, probably working, or clearly not, and that is a question you can actually answer.

If you want the textbook version of how marketing attribution is supposed to work, read the reference and then come back. This post is about what to do once the textbook meets a real business with real holes in its data, which is every business I have ever run.

How do I actually track where customers come from?

I run two cheap models side by side and let them argue. First-touch answers discovery: what introduced this person to me. Last-touch answers conversion: what was in front of them when they finally paid. Neither one is the truth. Together they bracket it, and the bracket is usually tight enough to spend against.

• First-touch for discovery. Credit the first source you ever recorded for a visitor. It over-rewards top-of-funnel channels like content and social, but it answers the question those channels are responsible for: did anyone new find me this month?
• Last-touch for conversion. Credit the final source before the sale. It over-rewards branded search and email, but it tells you what tends to close. Most tools default to some flavor of last-touch, which is why it quietly runs your reports whether you picked it or not.
• Tag everything you own. Every link you control gets a UTM: ads, newsletter buttons, partner links, QR codes. You will never tag dark social, but you can stop losing the channels you can tag through pure laziness.

There is no excuse for an untagged campaign link, so I keep a UTM builder one tab away and tag before I post. It does not fix attribution. It stops me from corrupting the little clean data I already have.

Just ask them: the how did you hear about us field

This is the highest-return attribution tool almost nobody bothers with, and it costs one input box. Add a single optional field at signup or checkout: how did you hear about us. Make it free text, not a dropdown, because a dropdown teaches people to click the first option and move on. The answers are messy, and they are also the closest thing to ground truth you will get, because they come from the one place no tracker gets to look: the customer's own memory.

I ran this and half the answers named things my analytics never once recorded. A specific podcast. A friend. A Reddit thread. Saw you everywhere for months. That last one is a multi-touch journey confessing itself in plain language. Self-reported data is biased and incomplete, and it is still the only source that captures dark social, so I weight it heavily and I read every single response instead of bucketing them.

Connect the money, not just the clicks

Here is the reframe that turned attribution from maddening into useful for me: stop attributing pageviews and start attributing payments. Traffic attribution asks which channel got a click. Revenue attribution asks which channel got paid. The second question is smaller, much harder to fool yourself about, and the only one your bank account has an opinion on.

This is the one place I will let my own product into the post, honestly. I built Conclick so you can connect Stripe, Paddle, Polar, Lemon Squeezy, or Dodo and tie each payment back to the source, campaign, and funnel that earned it. It does not make attribution certain, and I will not pretend it does. What it does is make the last mile, the part that touches money, stop being a guess, and it surfaces the gap between the revenue your processor reports and the revenue your analytics claims. That gap is where most founders are quietly lying to themselves.

What a good-enough attribution setup looks like

Put together, my entire attribution stack is four moves, and none of them need a data team or a six-figure tool.

1. Tag every link you own with UTMs, so the channels you can see stay clean instead of collapsing into direct.
2. Read first-touch and last-touch side by side, and trust a channel only when both roughly agree or when one is obviously carrying discovery.
3. Ask how did you hear about us in free text, and actually read the answers, because they are your only real window into dark social.
4. Tie revenue to its source, so the final call is made on money that cleared, not on clicks you are hoping converted.

Attribution is not a truth machine. It is a compass. Stop demanding GPS coordinates from a compass and it starts being useful.

Do all four and you will still be guessing. The difference is that you will be guessing with your eyes open, spending against directional truth instead of a fabricated decimal. I would rather be roughly right about where my money comes from than exactly wrong, and after enough launches, roughly right is the thing that compounds.