Lead Conversion and Offline Attribution
Decision question
A hospitality business was running Google Ads and Meta Ads, but a click or message did not show whether a person arrived at the location or generated revenue. The team needed to know which follow-up and audience approaches were associated with customers, not only platform engagement.
Records used
The work used existing prospect records, campaign activity, phone numbers collected in exchange for an offer, personalized offer codes, verified in-person redemptions, and recorded revenue.
These records were created during ordinary business operations. They were not collected through a randomized experiment.
Work performed
I defined the stages from campaign exposure to engagement, lead, verified visit, and purchase. I then introduced two related workflows.
First, I scored and ranked prospects from existing business records so the team could spend follow-up time on the strongest available leads. The ranking produced an ordered contact list. It did not determine why any individual converted.
Second, I set up phone-number capture and personalized offer codes. A customer had to present the assigned code to redeem the offer in person. That created a link between the follow-up record, the visit, and the revenue attached to the redemption.
Verified visitors then supported remarketing and lookalike audience construction. Campaign spending could be compared with recorded attributable revenue instead of platform engagement alone.
Observed results
The internal operational comparison recorded two changes:
- Average lead conversion increased about 30 percent after the prospect-ranking workflow was introduced.
- Cost per attributable dollar of revenue decreased about 70 percent after audiences based on verified visitors were introduced.
The reported percentages compare operating periods before and after the workflows. They describe recorded changes in this business. They are not estimates from a randomized controlled test, and other changes in targeting, creative, or customer demand may have contributed.
Business use
The system gave the team a way to distinguish a platform lead from a verified visit. Follow-up time moved toward higher-ranked prospects, while campaign review could use attributable revenue and later customer value rather than impressions alone.
What the result does not establish
The records support operational attribution and a before-and-after comparison. They do not isolate the incremental effect of each ranking signal or audience change. A stronger causal test would require a defined holdout group, stable offer conditions, and enough observations to compare conversion and revenue across groups.