Let statistical probability do the qualifying
Storytelling and emotional persuasion work. But the lever most service businesses ignore is the actual data behind the outcome — the statistical probability a buyer has of succeeding with you.
When someone sees a real success rate that matches where they actually are, they self-qualify in the first few seconds. When the number is nowhere near their reality, they skip. That's the point. You want the wrong-fit people to bounce before they burn your ad spend, your calendar, and your sales team's time.
If you've been in business for years, you're sitting on spreadsheets of this data already. Feed it into an AI workflow, slice it, and turn it into ad angles instead of sitting behind a desk for weeks doing it manually.
Break your offer into three tiers of success
If you have one service trying to serve everybody, the messaging can't align tightly with any single buyer. Break the outcome into three tiers: a beginner win, a middle-of-the-pack win, and a milestone win.
Each tier gets its own ad, its own angle, and its own success stat. A beginner in strength training wants to hit a 225 bench. A middle-of-the-pack business owner wants 50 sales a month. The milestone buyer wants a 700 deadlift or a million a month.
If I'm nowhere near a million a month and your ad opens with 'we took 15% of our clients to $1M/year,' I'm out in two seconds. Good. That ad wasn't for me. But it needs to exist for the person it is for.
Tier 1 (beginner): 50% achieve X in 3 months Tier 2 (mid): 30% achieve Y Tier 3 (milestone): 10% achieve Z
Lead with the stat, then answer the checklist
When someone has a problem, they build a mental checklist. Has anyone else gone through this exact thing? Did you help them? Are you fast? Any bad experiences? Do people actually like you?
Hit that checklist in the first 5–10 seconds — before the skip button lights up. Open with the probability of success or a time-frame stat, then stack a quick relatability line or testimonial.
Roofing: '80% of jobs we take last one week vs. the industry average of 1.5 weeks.' Cleaning: '90% of our clients rebook monthly for five years minimum.' Landscaping, coaching, whatever you run — pick the number that answers the checklist question your buyer is silently running.
0–10s: success stat + timeframe 10–20s: relatability / short testimonial 20–30s: what they DON'T have to do vs. competitors
Name the sacrifice they don't have to make
After the stat, tell them what stays the same and what they don't have to do. The dream outcome without A, B, and C — where A, B, and C are the things they'd be forced to do with a competitor.
This is the part most ads skip. People buy the absence of pain as much as the presence of results. Spell out the effort required, the sacrifice avoided, and what actually changes on the other side — status, ease, appearance, whatever their real driver is.
AI can pull these patterns straight from your client notes and testimonials if you feed it the raw data.
Run the evolution ad campaign — six sets, three variables
Don't run one ad. Run three ads (one per tier) with variations across hook, body, and CTA — six sets minimum. Especially on Meta with Andromeda, you need volume for the algorithm to learn.
Give it 1–2 weeks before judging. You don't want to react to regression to the mean. Once winners emerge, keep them running untouched. Kill the zeros. For the middle performers, mix and match: take the winning hook, the winning body, the winning CTA, and rebuild the losing slots from those parts.
That's the evolution — you're not rewriting from scratch, you're recombining what already converted.
Ad 1: +5 → keep Ad 2: +7 → keep (winner) Ad 3: +3 → rebuild Ad 4: +5 → keep Ad 5: +4 → rebuild Ad 6: +0 → kill
Close the loop on attribution or the algorithm lies to you
Here's what breaks most ad accounts: you're only feeding form fills back to Meta and Google. The algorithm thinks a form fill is the goal, so it brings you more form fillers who never buy.
You have to send closed deals back too — and disqualified leads, and no-shows. That's closed-loop attribution. Google needs the closed deal within 90 days or it won't influence the algorithm. Meta's pixel holds data up to 180 days, but the most recent 2–4 weeks are what's actually driving who sees your ads right now.
If you had three great years and one bad year, the algorithm doesn't care about the great years. It's optimizing off last month. Feed it the right signal or it will keep feeding you the wrong leads.
Send back to Meta/Google: - Closed deals (within 90 days) - Disqualified leads - No-shows - Not just form fills