Connecting Paid Campaigns to Verified Visits
A lead only counted once the business could tie it to a visit and a sale.
- Client
- A cafe that later became an event lounge. Independent work.
- My role
- The only analytics and marketing person. I set up the tracking, ran the Google and Meta campaigns, and reported to the owner.
- Result
- On lead campaigns, the ad spend behind each dollar of tracked sales fell about 70%.
- How it was measured
- Before and after the audience change, counting only sales tied to a campaign by a personal offer code. Not a randomized test.
Result: on lead campaigns, the ad spend behind each dollar of tracked sales fell about 70%, after I rebuilt the ad audiences from customers who actually walked in.
The problem
A cafe, later an event lounge, was running Google Ads and Meta Ads, and the platforms reported clicks, messages, and leads. None of that said whether a person ever walked through the door or spent money. The owners wanted to know which campaigns brought in paying customers.
What I built
I started by defining the funnel: exposure, engagement, lead, verified visit, purchase. Then I closed the gap between the ad platforms and the till.
Dated flyers first. The earlier owner spent about $1,000 to $2,000 a month on printed flyers with no tracking. I dated each flyer and tied it to a free drink with a food purchase, so every redemption traced back to a campaign and gave a cost per acquisition. It was the business's first campaign-level view of offline return, and it showed print wasn't paying for itself.
Phone numbers and personal codes. For paid campaigns, customers gave a phone number for a discount and had to show a personalized code to redeem it in person. Each redemption tied a campaign to a verified visit and the revenue attached to it.
Audiences rebuilt from real visitors. Verified visitors fed remarketing, which shows ads again to people who already engaged, and lookalike audiences, which find people who resemble past visitors. I ran two funnel shapes: a direct path from ad to visit, and a two-stage path that converts people to Instagram followers first and to visits second. Lead campaigns cost a lot to acquire, so I ran customer lifetime value analysis, what a customer spends over time, on each funnel stage to see which ones paid for themselves.
How the 70% was measured
The measure is cost per attributable revenue dollar: lead-campaign spend divided by the revenue from the verified visitors those campaigns brought in. A visit counts only when the customer redeems the personal code tied to a campaign. I compared that figure for lead campaigns before and after the audiences were rebuilt from verified visitors. It's an operating comparison rather than a randomized test, and it can only see customers who redeemed a code.
What the spend data showed
With a new audience, return on ad spend (sales per dollar of ads) started near 3x and drifted to about 2x as the campaign aged. At that scale, broad reach to new people beat a large retargeting campaign, because the local market was far from saturated, and retargeting paid off only in small doses. The flyer test also settled an argument with an owner who was set on print: the flyers weren't cost-effective.
How the owners used it
The owners could finally tell a platform lead from a paying customer, and campaign reviews ran on attributable revenue and customer value instead of impressions.
I was the business's only analytics and marketing person. I reported to the owner, briefed the creative contractors and directed their work, and when the owners disagreed, I settled it with a test and a range of likely outcomes instead of an opinion.
I also tied marketing to capacity. I forecast visits by hour as a likely low-to-high range, so the owners knew how much food, stock, and staff a given spend would need. I used those forecasts to temper a new owner's plan to spend more. Demand past capacity wastes the budget and hurts word of mouth.