The four pricing models
Nearly every SEO quote you'll receive uses one of four structures. Each has a legitimate use and a characteristic failure mode:
| Model | How it works | When it fits | Failure mode |
|---|---|---|---|
| Monthly retainer | Fixed fee for an ongoing program of work | Sustained visibility — the normal shape for local SEO | Retainers that drift into "maintenance" where nothing ships; demand an itemized monthly deliverable list |
| Project | One-time fee for a bounded piece of work | Site rebuilds, technical cleanups, profile overhauls | The gains erode after the project ends if nobody defends them |
| Hourly | Billed per hour, usually for consulting | Audits, second opinions, in-house team support | Misaligned incentives on open-ended work — slow burn pays better than fast results |
| Performance | Pay per lead, per ranking, or per outcome | Rarely — read the fine print twice | Lead-quality games (billing for spam calls), rank "guarantees" on keywords nobody searches, and lock-in via assets the vendor owns |
Performance pricing deserves the extra warning. "You only pay for results" sounds like aligned incentives, but the vendor defines the result: a guaranteed first-page ranking is trivial on a keyword with no searches, and a per-lead price is only as good as the vendor's incentive to filter out junk calls — which is zero. Some per-lead operations also build the assets (site, phone numbers, profiles) under their own ownership, so leaving means starting over. If a performance deal doesn't survive the questions later on this page, it wasn't a deal.
What actually determines cost
- Market competitiveness. Outranking thirty funded roofing companies in a major metro is a different project than standing out among four in a small town. Same trade, different budget.
- Service-area size. Every additional city you want to win in is additional pages, additional Maps presence, and additional review gravity to build. Cost scales with geography more than with company size.
- Starting condition. A technically sound site with a claimed, active Google Business Profile starts the race mid-track. A neglected site and an unclaimed profile mean months of structural work before growth work begins.
- How many surfaces you're contesting. Rankings alone is a smaller scope than rankings plus Maps plus AI answers. Most home service companies end up needing all three, because that's how homeowners actually search now — but scope, and price, should name the surfaces explicitly.
Is "AI SEO" or "AEO" priced separately?
A growing number of proposals now list "AI SEO," "AEO," or "GEO" as a separate package or an add-on line item. Read those carefully. Most of the work that earns AI-answer visibility — content that actually answers questions, a complete Google Business Profile, consistent business data, a real review base — is the same work that earns rankings and Maps presence. There is genuinely new work in the mix (measuring what assistants say about you, structuring content so an answer engine can lift it cleanly), and that new work justifies some scope. What it doesn't justify is paying twice for the same deliverables under two names.
As of September 2026 there is no standardized market rate for AI-visibility work — the category is too new — which is one more reason to evaluate any quote by its itemized deliverables and the break-even math below, not by what the package is called. If a proposal lists SEO and AEO as separate packages, ask which deliverables appear in both columns.
The evaluation math
Whatever anyone quotes you — us included — the decision comes down to one formula: break-even jobs per month = monthly cost ÷ gross profit per job. If the break-even number looks easily achievable from better visibility in your market, the quote is rational. If it doesn't, walk.
Worked example — hypothetical numbers, run your own
Say you run an HVAC company where the average system replacement is $9,000 at a 35% gross margin — $3,150 gross profit per job. Against a hypothetical $3,000/month retainer: $3,000 ÷ $3,150 ≈ 0.95, so the program breaks even at one additional replacement per month. Every call beyond that is margin. Now the honest flip side: if your average job is a $250 repair at the same margin ($87.50 profit), that same retainer needs 35 extra jobs a month to break even — a much harder case, and a sign that scope (or the service mix you're optimizing for) should change. The formula matters more than our example: put your job value and your margin in it before signing anything.
Questions that expose a weak proposal
Ask these of every vendor. Confident, specific answers are the tell — in both directions:
- Who owns the site, content, and profiles if we part ways? The only acceptable answer is "you do, all of it." Anything else is lock-in with extra steps.
- What exactly ships each month? Not "ongoing optimization" — a list. Pages, fixes, reviews generated, reports. If they can't name deliverables in advance, they can't be held to them.
- How do you measure AI visibility? If the answer involves a proprietary score they can't show the method for, it's marketing. A real answer names the prompts, the tools, and the cadence — ours is published.
- Which of my competitors do you already work with? An agency serving two roofers in one city is playing both sides of a zero-sum game. Ask directly; watch for hedging.
- What happened with your last client in my trade? Listen for specifics about the work and the numbers — and for whether they'll honestly describe one that didn't go well.
- What would make you tell me to spend less? A vendor with no answer has no scope discipline. There are markets and starting points where the right program is smaller — a good one will say so.
The terms worth insisting on — from anyone
No package grid on this page, for the reasons above. What it can give you instead is the answer key — the terms that count as good answers to the questions above, whoever is quoting you:
- Scope follows evidence, not a package name. Your baseline across the three surfaces — which you can establish yourself today with the free 25-prompt audit — determines what actually needs doing, and any quote should be itemized by surface and deliverable against it.
- No long lock-in. An initial term long enough to do real work is fair; a multi-year contract that outlives the vendor's motivation is not. Month-to-month after a short initial term is a reasonable ask in this industry.
- You own everything that gets built. Site, content, profiles, data — leaving should cost you nothing but the relationship.
- No working both sides of your market. Local search is zero-sum. Ask directly whether they serve your competitors in your service area, and treat hedging as an answer.
Take the break-even math with your numbers, the itemized deliverable list, and measurement you can replicate yourself into every sales conversation. A vendor with nothing to hide won't flinch at any of it.