How to Price a White-Label AEO Retainer Without Capping the Work

You can sell a white-label AEO retainer that clients can inspect and that still holds its margin at client ten. First, stop pricing it like the old keyword package: 25 tracked prompts on Basic, 100 on Pro, more if the client pays up. I used to sell retainers that way. I was wrong about what the client was buying.

The client does not care how many keywords you watched. They care what shipped: pages published, technical blocks removed, citations pursued in ChatGPT and Google AI Overviews, with a log to prove the work. Charge per domain, sell executed work the client can verify, and treat the platform cost as a predictable line item per domain. That is a retainer you can explain at five domains and at fifteen.

Why does the old tier sheet fail?

I keep seeing the same Bronze, Silver, Gold table repackaged for AI search. Twenty-five prompts on Basic, 100 on Pro, 350 if you really pay up. It looks familiar because it is the old PPC tier sheet with keywords crossed out. The caps underneath tell the story: Profound Starter at $99 a month is ChatGPT-only with 50 prompts, and Growth costs $399 for three engines and 100 prompts. Peec runs $95 for 50 prompts, $245 for 150, and $495 for 350. Sell the cap as your package and you spend the engagement arguing about how many prompts the client gets to watch.

Clients do not renew because you watched more prompts. They want to see pages, technical fixes, and whether they are being cited in the answers their buyers read. Measurement tools run from $29 for Otterly Lite to $250 for Scrunch Core, and Semrush charges about $99 per domain for 25 prompts. The software that measures a citation does not create it. Content, schema, reviews and PR do the unglamorous work; together, they account for 50 to 70% of a working budget. Price the watching and you cap the wrong deliverable. Price the domain and keep a work log, and the client can see what the retainer bought.

Step 1: What are you actually selling?

A visibility report tells the client where they stand. Shipped work gives them a reason to expect that position to change. I learned this running PPC: nobody kept paying for a search-term report; they kept paying when wasted spend dropped and pipeline rose. AEO has the same distinction. Promise an AI ranking and you may spend month three explaining variance. Promise defined work and a log of what went live, and the client can check delivery without taking your chart on faith.

Put the scope in verbs. Depending on the domain, that can mean:

• Optimising entities across Wikidata, schema and sameAs links.
• Publishing extractable definitions and comparison tables.
• Handling llms.txt and AI-permission setup where relevant.
• Pursuing trusted-source citations through editorial and forums.
• Tracking prompts across engines and defending visibility in AI Overviews.

That is the white-label AEO deliverable list I would use to test a proposed package. It does not mean promising every item for every domain, or promising a citation because you published a page. It means naming the work, the access it requires, and what the client will see in the log. Skip anything that reads like repackaged SEO with an AI label. Below about $1,500 a month retail, that is usually what it is; boutique AEO retainers cluster at $3,000 to $10,000 a month for mid-market. Sell the work log, not the score.

Step 2: Which software can deliver that work?

Two kinds of tools carry the same AEO label. Monitors watch prompts and chart who got cited. Execution platforms help ship the work intended to earn visibility: pages, technical fixes and references. Profound caps content at zero articles on Starter and three per month on Growth, Peec offers an actions roadmap but no writing, and AirOps keeps its Quill writing agent behind the Pro jump. Package a monitor as a full retainer and you still owe the client the labour. It comes out of strategist hours or contractor spend. Either way, it comes out of margin.

Then ask what white label actually means. Geoptie and Semrush offer branded PDFs; SE Ranking puts the platform on your domain through an extra Agency Pack; Otterly leans on a Looker Studio workaround; Scrunch provides API outputs only; and Profound and AthenaHQ include no white label on agency plans. A logo on a PDF is not the same thing as a delivery workflow your agency can own.

My shortlist has four tests:

1. Can reports go out under your brand?
2. Does the cost per domain stay predictable as you add clients?
3. Can you track the prompts you need without overages dictating the scope?
4. Can work reach the CMS without sitting in a ticket queue?

I run the last test first. If the tool cannot help you ship, treat it as a report, not a delivery engine. The agency AEO software ranking walks through that monitor-versus-execution split at client ten, when per-workspace pricing starts to bite. Buy the layer that matches what you promised to deliver.

Step 3: What should you charge per domain?

Total first: a sellable single-domain AEO add-on retails for $1,800 to $4,500 a month; competitive work retails for $4,000 to $9,000. Wholesale delivery runs $600 to $1,500 a month for a single brand and $1,500 to $3,500 for competitive work. The gap is not profit. It still has to cover your strategist time and any costs outside wholesale delivery.

The setup I see hold is a $2,000 AI-visibility line on an existing SEO retainer against about $900 in wholesale delivery, supported by work and citation checks the client can see each week. Here are the listed figures before the less visible costs enter the room:

Do not compare a tracker subscription with wholesale delivery as if they buy the same work. The tracker price excludes the publishing, fixes and citation effort your retainer promises. Seats, workspaces and add-ons can obscure the monitoring bill, too: ten clients tracking 40 prompts each cost $399 a month on Geoptie Enterprise versus $4,089 on Profound Agency Growth, before add-ons such as $45 per extra Semrush user or $439 for the Otterly Claude add-on. Those are monitoring costs, not a quote for delivery.

The cheapest mistake is calling a tracker a retainer. The most expensive is signing per-workspace monitoring and discovering at client eight that each new domain eats more margin than you allowed for. I price per client account on an execution platform like groas for agencies, where content, fixes, citation work and weekly tracking can ship under the agency’s brand. Keep the platform as a clear line in the model. Do not bury it inside a prompt allowance the client never asked to buy.

Step 4: How do you onboard a domain without a month of meetings?

Do not start with a 40-page audit. Start with access, a prompt set and a baseline you can defend in month three. I ask for three things on day one:

1. CMS and Search Console access.
2. The five pages the client says drive revenue.
3. The ten buyer questions the sales team hears most often.

Those questions become the first prompt set, sampled across ChatGPT, Perplexity, Gemini and Google AI Overviews. Perplexity and ChatGPT can pick up new content in days while AI Overviews lag a full crawl cycle, so baseline each surface before you change anything. Otherwise, month-three movement becomes a story you remember rather than a comparison you can show. The walkthrough in onboarding a white-label AEO client in an afternoon uses the same order: route the domain, run the gap audit, publish the first fixes, then let the log do the talking.

Keep technical access narrow and reversible. A connection for the ad accounts and site, working across platforms or CMSs in a couple of clicks, beats a week of developer tickets. Check the basic path first: can an engine reach the page, parse the answer block and attribute it? If crawlers cannot reach a money page, or its key comparison sits inside JavaScript they do not render, more prompt tracking will not fix it. Ship the access fix and one extractable block per money page in week one. Baseline first, then ship something the client can inspect. That early work helps the retainer survive a quiet first month.

Step 5: What goes in a report clients will keep reading?

Set the clock honestly. First citation movement can take four to eight weeks, meaningful share three to six months, and compounding six to twelve. I tell clients to judge at month four and put a 90-day checkpoint in writing. That gives us a point to examine the work and the direction without pretending week two is a final verdict.

The weekly report needs three things:

1. What shipped, with the reason for each action.
2. Which engines moved against the baseline.
3. Which tracked prompts were won and lost.

No share-of-voice dial without the work behind it. If a prompt moves, the client should be able to see what changed; if it does not, the log should still show what you tried. The margin math behind that cadence is laid out in how to hold margin at ten client domains: one predictable platform line per client account, a few strategist hours for judgment, and a log the client can forward to their boss. Report decisions and executed work, not just a prettier chart.

When should you skip the AEO retainer?

Skip it when the foundation will eat you alive. I turn down three kinds of domains:

• The site you cannot change: no CMS access, a dev queue measured in quarters, or key pages trapped in JavaScript the engines do not render.
• The business with little to cite: no reviews, no case work, no named customers, in a category where ChatGPT and Perplexity quote sources with history.
• The client who needs signed deals in 30 days: they hear AI visibility as immediate lead gen, then judge the work before it has had time to appear.

That last client can churn in week six no matter what you ship. Perplexity may reflect a new page in days while AI Overviews waits on a crawl cycle. Sell them PPC. Come back to AEO when you can publish weekly, fix access without a ticket, and point to something real worth citing. A no now protects the margin on the domains you can actually serve.

For the domains that pass, keep the arithmetic plain. At $2,000 a month retail against roughly $900 in wholesale delivery, you have about $1,100 per domain before strategist time. Five domains leave around $5,500 a month of spread; ten leave around $11,000; fifteen leave around $16,500. That is spread, not take-home profit. The model works when the report template, prompt sampling and publishing motion repeat while the cost of each added domain stays predictable. The prompt-capped monitor route bends the other way when workspaces, seats and overages stack. The margin guide for ten client domains describes the operating split: two to three strategist hours per client for judgment and client-facing time, repetitive work handled by the execution layer, and actions logged for renewal.

Pick one client domain today and run the baseline before you sell anything. Pull ten sales-call questions, check them across ChatGPT, Perplexity, Gemini and AI Overviews, and capture who gets cited and which client pages could have answered but did not. That sheet gives you a scope, a price justification and a month-four scorecard. Then choose an execution platform that ships under your name rather than a tracker that only watches. groas offers full-service white label for agencies: content, technical fixes and citation work per domain, tracking across Google and AI engines, and weekly reports under your brand. Start with your first five client accounts at $299 a month each, per service, backed by a 7-day money-back guarantee, keep the log, and price the next domain for the work you will do.