How to tell which marketing is actually working

Almost every service-business owner spends money and time on marketing (ads, posts, a booth at the local event, a referral perk, the hours poured into social) and almost none can say with confidence which of it is actually working. The spending is concrete and the return is a fog. So budgets get set by habit and gut: you keep doing what you've always done, double down on whatever felt good lately, and quietly suspect that half of it is wasted without knowing which half.

The reason this stays foggy isn't that the answer is unknowable. It's that most owners measure marketing with the wrong ruler, counting the activity it generates instead of the clients it produces, and never close the loop between a channel and the money it actually brings back. Closing that loop doesn't require an analytics degree. It requires changing what you count.

Stop counting activity. Start counting clients.

The seductive marketing numbers are the ones platforms hand you for free: impressions, reach, likes, followers, clicks. They feel like progress because they go up when you post and spend. But none of them is money. A post can reach ten thousand people and book nobody; an ad can rack up clicks and produce only tire-kickers. These metrics measure how loud you were, not whether anyone walked through the door because of it.

Even “leads” can mislead. A channel that generates lots of enquiries but few bookings, or bookings that never return, isn't working. It's manufacturing activity. The only marketing metric that ultimately matters is paying, returning clients, and the revenue they bring over time. Everything earlier in the funnel is a proxy, and proxies are exactly where marketing money goes to hide.

The simplest attribution beats the fanciest: just ask

Here's the low-tech method that outperforms most analytics setups for a small service business: ask every new client how they found you, and write it down. “How did you hear about us?” at booking or on the first visit, captured somewhere you can count it later. It feels almost too simple to matter, but over a few months it builds the single most valuable marketing dataset you have: a direct, human line from each client back to the channel that produced them.

This works precisely where digital attribution falls apart. A client might see your Instagram, get recommended by a friend, walk past your sign, and finally book after a Google search, and no tracking pixel untangles that. But ask them and they'll tell you what actually tipped them: “my sister comes here.” That's the channel that did the work. Word of mouth and referral, the highest-value sources in most service businesses, are nearly invisible to analytics and obvious the moment you ask. Don't overlook the method just because it isn't automated.

Judge each channel by the clients it produces

Once you're tracking where clients come from, you can do the thing that actually answers the question: compare channels by the quality of the clients they bring, not the quantity of attention they generate. Some channels reliably produce clients who book a premium service, rebook, and refer. Others produce clients who come once for a deal and vanish. On a reach-and-clicks dashboard those can look identical or even favour the worse one. Measured by the clients themselves, they're worlds apart.

So look at each source through the lens of the clients it sends: Do they come back? Do they pay full price or only book promotions? Do they refer others? A small channel that quietly sends you loyal, high-value clients deserves more of your budget than a big one that sends churn. This is the difference between spending where the noise is and spending where the value is, and you can only see it once you're counting clients by source instead of activity by platform.

Give it a fair window, then check what you predicted

Marketing is lumpy and slow, so timing trips people up. One strong month doesn't prove a channel works and one weak month doesn't prove it's dead: service-business marketing has long, uneven lags between effort and result. Before you judge a channel, give it a fair window and enough volume to mean something, and resist the urge to kill or crown it on a single data point.

Then apply the discipline that turns marketing from guesswork into learning: before you commit to a channel or a campaign, write down what you expect it to do, not “more awareness,” but something checkable, like “I expect this to bring in fifteen new clients over the next two months, and I expect a third of them to rebook.” When the window closes, compare. Did it deliver the clients, and did those clients come back? That comparison is the whole game. It's the same method behind knowing what's actually working in your business: predict, run one thing, and check whether reality agreed, applied to the budget that's easiest to waste on faith.

What to do this week

1. 01Start asking every new client how they found you, and record it somewhere you can count later. This one habit builds your best marketing dataset.
2. 02Stop reporting on reach, likes, and clicks. Replace them with one number per channel: paying clients produced.
3. 03For your top channels, look at whether the clients they send actually rebook and pay full price: judge by client quality, not volume.
4. 04Before your next campaign, write down what you expect it to deliver, give it a fair window, and put a date on the calendar to check who actually came back.

The part that quietly breaks down is the remembering: tying this month's clients back to the channel that produced them, and recalling three months later whether that campaign's clients ever returned at full price. Keepr keeps that thread for you. It reads your bookings, payments, and the sources you capture, connects new and returning clients back to where they came from, and remembers whether a channel produced loyal clients or one-time deal-seekers, so the next budget decision starts from evidence instead of the gut. The hard part of marketing was never spending the money. It was knowing, afterward, which of it worked.