Seven checks, in the same order, on every proposal.

A fixed rubric beats skimming for things that feel wrong, because the clauses that cost the most money are the ones that read as boilerplate. The same order applies to GEO and AI-search proposals as to SEO ones.

Check my proposal for all 18 flags What the report contains
  • $199
  • verdict in 24 business hours
  • I review every proposal myself
  1. 01

    Parse the proposal

    I start by pulling every real promise out of the proposal and writing it down as a list: the price, how often you are billed, the term and notice period, each named deliverable with its quantity, and who owns what at the end. Slide decks and email threads get the same treatment. Six pages of pitch can shrink to a handful of commitments. That list is then held against what SEO agencies do when the work is real: counted deliverables, dates and named people.

  2. 02

    Check it against 18 red flags

    Then I work through 18 red flags, in the same order every time, rather than skimming for things that feel wrong. The flags that cost the most money are the boring ones, like who owns your content and accounts, a silent auto-renew, or no way to cancel at all, because they read like boilerplate and nobody argues with boilerplate. Some flags fire on what is missing. No termination clause? That gets flagged too. Each flag I raise gets a severity from 1 to 3, the exact quote and where it sits, and what it costs you in plain words. If the proposal sells GEO, AEO or "AI visibility", I run five AI-search flags as well: what is genuinely new versus SEO with a new label, deliverables nobody can measure, promises to rank you in ChatGPT, and AI tracking that is manual or never defined. The rubric is built from the wording that SEO scams leave on paper, so it looks for phrases, not impressions.

  3. 03

    Price check

    I compare the monthly fee with what a similar scope usually sells for, and with what the listed work would cost to do honestly. This is where inflated keyword volumes and bulk AI content tend to show up: a retainer justified by global search volume presented as local demand, or fifty articles a month that cannot be reconciled with the team described three pages earlier. AI-search work has no public price survey yet, so the report says so and prices it by the hours it buys instead of inventing a benchmark. The benchmark is public survey data on SEO services cost, not a guess.

  4. 04

    Site reality check

    I open your live website and look at what it ranks for now, what is clearly broken, and whether the plan deals with any of it. A proposal that never mentions your actual site was written before anyone looked.

  5. 05

    Check the agency's own claims

    Proposals lean on case studies and reviews, so I check those too, using public evidence only. For up to three cited case studies: is the client a real business, does its site rank for what is claimed, does the timeline fit the domain age and the Wayback Machine record, and is there any visible link between the agency and that client. Each is rated verified, plausible but unverifiable, or doubtful. Then I read the agency's reviews across Google, Clutch, Trustpilot, the BBB and similar sites for bursts, one-review accounts, copied wording and ratings that disagree between platforms. The result is a short read: consistent, mixed or concerning. I describe what the evidence shows. I never accuse anyone, and a pattern read is not proof.

  6. 06

    Verdict and risk score

    I give one verdict: Fair, Overpriced or Walk Away. I build the 0-100 risk score from the flags raised and their severity, so two audited proposals sit on the same scale. Stacking matters more than any single flag: a long term plus a silent renewal plus no exit clause is a trap regardless of how each one scores alone, and when that combination appears it is named explicitly. Most of those traps sit in the SEO contract terms rather than in the pitch.

  7. 07

    Negotiation script

    The last section is an email addressed to the agency, written for you to paste and send. It asks for the specific amendments that would make the agreement signable, in the order they matter, in language that is firm without being hostile. Where a flag has a known fair version, the script asks for that version by name rather than complaining in the abstract.

Usually 6 to 10 pages, depending on the proposal. Page one answers the question.

A PDF in seven fixed sections, its length set by how much the proposal contains. No dashboard, no login, no video walkthrough you have to schedule.

Verdict and risk score
Page one. One of three verdicts, the score out of 100, and a short paragraph saying why. If you read nothing else, this is the part that answers the question you paid for.
The top 3 issues
The three findings that carry the most money or the most lost control, each with the quoted line, its location in the proposal, and what to do about it.
Full flag table
Every flag raised, with its ID, severity, the quoted evidence, and the fair version. Flags that were checked and not raised are listed too, so you can see the proposal was read in full.
Price assessment
What you are being asked to pay, what that scope typically costs, and what the total commitment is over the full contract term rather than per month.
Site notes
What was observed on your live site, and whether the proposal reflects it.
Case studies and review signals
Each checked case study with its rating and the evidence behind it, the review read with the specific signals found, and the question to ask the agency where something could not be confirmed.
Negotiation script
The last page. It's written so you can copy it and send it the same day.
What the report contains

What I promise, in writing.

24 business hours from receiving your proposal
The clock starts when I receive your proposal through the intake form, not when payment clears. Weekends and US federal holidays do not count. Submit on a Friday afternoon and the report is due by Monday afternoon. The agency checks are capped so they fit: up to three case studies, and the main public review sites. Anything that cannot be confirmed in that time is reported as unverifiable, not guessed.
Missed deadline means a full refund
If the report is late, the $199 comes back and you still get the report. No claim form, no argument.
I read every proposal myself
AI helps me with the research and the first draft, but I read every proposal myself and write the verdict. Every flag, quote and judgement in the delivered report has been checked by me before it is sent. The verdict is my opinion, signed.
Seven days of follow-up
Reply to the delivery email with questions about anything in the report for 7 days after delivery. I answer by email, at no charge.
A free re-audit of a revised proposal
If you send the negotiation email and the agency comes back with a revised proposal, I re-audit it once, free: same agency, same engagement, within 7 days of your first report. You get what changed, which flags are fixed, which are still open, and an updated verdict. The re-audit is included free, carries no late-delivery refund, and has no cash value. A second revision is a new audit.
Your proposal is deleted after 7 days
The files you upload and any text you paste are deleted from storage automatically after 7 days. My working copy and your report are deleted 90 days after delivery, or sooner if you ask. Anonymized benchmark data (price, term length, flags raised) is kept so the rubric keeps improving.

Before you sign, get it read by someone with nothing to sell

A flat $199. One proposal, read line by line. A written verdict within 24 business hours.

Check my proposal for all 18 flags
  • Flat $199, paid once
  • Report within 24 business hours or a full refund
  • Full refund if you never send the proposal
  • I don't sell SEO