Insights
Attribution, for businesses that close offline.
What we have learned reconciling closed sales against lead sources — the methods, the failure modes, and the numbers that survive a finance review.
Your contact form counts submissions. Count the ones that paid
A website form reports every submission as a lead: the bot, the typo, the fake email and the same person twice. Only a fraction of those rows carry a real person you could ever find in your sales records. Here is how to clean form leads, which fields make them matchable later, where the entries actually end up, and how to judge forms by sales.
Your brand campaign has the best ROAS in the account. Ask who sent the searcher
A search ad on your own business name will nearly always post the best return in paid search, because the person typing your name had already chosen you. A truck wrap, a referral, a billboard or last year's job sent them. Here is when brand campaigns still earn their keep, how to judge paid search with brand taken out, and how to test what the brand ads add.
Your UTM tags work. The source just never reaches the sale
UTM parameters tell your analytics where a visit came from, and nothing else. The source falls off on redirects, between pages and domains, at forms with no hidden fields, on every phone call and whenever staff retype a lead, and inconsistent naming splits what survives. Here is a setup that carries it to the CRM, and the backstop for when it does not.
A promo code proves somebody had the code, not that the ad sent them
Promo codes look like the cleanest attribution there is for phone and in-store sales: the code names the campaign, and the invoice records the code. In practice codes leak to coupon sites, get applied by staff to close a sale, go unasked for on the phone, and cost real margin every time. Here is when a code is good evidence, when it is not, and how to check codes against sales matched by phone and email.
Retargeting ads take credit for buyers who were already coming
Retargeting shows ads to people who have already visited, so its buyers were close to buying before the first impression. View-through credit, last-touch reporting and overlap with the channel that found them all inflate what it claims. Here is how to see which leads retargeting actually started, what experiments say it adds, and how to report it honestly.
Snapchat Ads for local businesses: count the customers, not the swipes
Snapchat reaches more than half of American adults under 30, and its lead forms arrive prefilled and cheap. For a med spa, an orthodontist, a dealership or an apartment community that closes by phone or in person, neither the swipes nor the cost per lead says whether anybody bought. Here is what Snapchat counts, why younger buyers break the usual match, and how to judge the channel by matched sales.
A Pinterest save is a plan, not a sale. Count the sales it turned into
People save a kitchen, a garden or a wedding look on Pinterest months before they hire anybody. Pinterest's reporting looks back 60 days at most, and a save carries no name, phone or email. Here is what Pinterest's metrics and attribution windows actually measure, how to capture leads that can be followed to a sale, and how to price Pinterest by the customers it produced.
Apple Ads counts installs. Your app's sales close somewhere else
Apple's App Store ads, called Apple Ads since April 2025, report taps, installs and cost per install, and attribute them in ways designed so you never learn who installed. For a booking, quote, dealership or clinic app, an install is a door opening. Here is what Apple reports, what its privacy design deliberately withholds, and how to follow app sign-ups to the sales that close at your counter.
Cheap clicks from inside other apps: judge in-app ads by the sales
Ads shown inside other people's apps, through Google's Display Network and AdMob, Meta Audience Network, AppLovin or Unity, produce the cheapest clicks most local businesses ever buy. Some are thumbs that missed a button, some are fraud, and a few are customers. Here is how to tell them apart with the one record that cannot be faked: who paid you.
The call nobody answered is still on your marketing bill
For a business that sells by phone, a missed call is a lead that was paid for and then dropped. It also quietly rewrites channel ROI: the channel whose callers ring after hours looks weaker than it is. Here is how to measure missed calls and callback speed from a call log, and how to report them without inventing revenue.
A chat without a phone number or an email is a conversation, not a lead
Your chat widget talks to buyers every day, and your attribution report has never heard of most of them. Visitors stay anonymous, chat tools record the page but not the ad, and the best conversations get retyped into the CRM by a salesperson with no source at all. Here is why chat leads vanish from revenue attribution, and the four changes that make them count without strangling the chat.
Influencer marketing for businesses that close later: count customers, not codes
Creator deals are judged on followers, engagement rate and code redemptions, and none of them is a sale. For a business that closes on a call, at a consultation or after a quote, codes leak and go unasked for, and platform reports count events rather than customers. Here is how to give every creator a door you can see through, match what comes through it to closed sales, and be honest about the rest.
Stop guessing which ads pay off.
Two exports and a few minutes. 14 days free, no card, nothing to install.