How much do AEO services cost

Answer Engine Optimization services cost a stack, not a screenshot: a one-time audit, a monthly retain (if you hire ongoing labor), the DIY hours you will still spend internally, and the tools that log prompts. Anyone selling you one monthly number without those four lines is selling a costume. The public ranges below are dated planning bands. They are not Spurlock Studios package prices, not a day-rate, and not a quote.

This spoke sits under the Answer Engine Optimization playbook. Whether the spend is even rational is when AEO is worth it. Who should do the work is DIY AEO vs hiring. What belongs in the audit vs the monthly object is AI visibility retainer vs audit. This page is the invoice math. I have been SEO certified since 2021. I will not invent a studio SKU to make the spreadsheet prettier.

The short answer

• Four lines, one total. Audit + retain + DIY hours + tools. Add them. Do not pick the smallest sticker and call it “AEO.”
• Agency bands are hearsay. Digital Elevator’s July 23, 2026 guide is the dated public table I will cite. It is still not a survey of invoices, and it is not our rate card (Digital Elevator AEO/GEO pricing guide).
• Tools are measurement. Otterly, Profound, and Semrush publish stickers. Those stickers do not rewrite a money page.
• DIY is payroll. Clock hours × your loaded rate. A BLS wage is a labeled proxy, not a founder’s real cost.
• Buy the object you are missing. No baseline → audit. Ranked list and nobody ships → retain. Eligible pages and a named owner → DIY plus a cheap log, not a thin retainer.

Line	You are buying	Planning input (not a quote)	Skip when
One-time audit	Dated baseline + ranked 30/60/90	A project quote for that artifact	The vendor cannot name five URLs
Monthly retain	Recurring labor that ships HTML	Digital Elevator mid-market band, then their SOW	You still cannot describe current state
DIY hours	Your people on a frozen prompt panel and tickets	Hours × loaded payroll (or a labeled BLS proxy)	Nobody will put the hours on a calendar
Tools	Prompt logging, not citations	Vendor pricing pages, re-fetched this quarter	The tool is the entire “program”

A cheap line that never ships HTML is a screenshot subscription. Pay for the stack you will actually run.

Why is there no honest single AEO price?

Because “AEO services” is four different purchases wearing one acronym. An audit is a diagnosis. A retain is a month of shipping. DIY is a payroll envelope. A tool is a meter. Averaging those into one “typical cost” is how a $99 monitoring plan and a $10,000 content-plus-PR retain become the same slide.

Google still does not sell an AEO SKU. A supporting link in AI Overviews or AI Mode has to be indexed and snippet-eligible in ordinary Search. There is no extra technical bar and no extra markup track (Google Search Central: AI features; Google’s generative-AI optimization guide, May 2026). You are not buying a Google product. You are buying labor and logging against a surface Google already defined.

Object someone called “AEO”	What you actually got	Why a single price lies
One-time audit	Baseline, eligibility, ranked tickets	Done when the readout ships
Monthly retain	Panel + entity hygiene + page jobs	Recurs; citations decay
Productized “AI SEO”	Template FAQ and a dashboard login	Scope is the template, not your URLs
Tool seat	Prompt tracking	Zero HTML
Classic SEO renamed	Rank reports with an Overview screenshot	You paid twice for the same hours

Decision list — you do not have a price yet when:

1. The proposal does not say audit vs retain vs tool vs staff hours.
2. “Implementation included” has no URL list and no stop date.
3. The number is “like SEO, plus 20%” with no jobs named.
4. The only artifact is a vendor AEO grade.
5. You cannot point at the internal hours that will still happen.

Price the object. Then compare quotes. A median of unlike objects is fiction.

What four lines make up AEO cost?

Audit, retain, DIY hours, tools. Write them on one sheet. The total is the program. Any line you hide still gets paid — usually by the founder on Sunday.

The cousin posts own the jobs. This table owns the bill.

Line	Cadence	Done looks like	Hidden cousin
Audit	Once (then stop)	Eligibility + frozen prompt set + ranked tickets + 30/60/90	A six-month “audit” that is a retain
Retain	Monthly	Shipped HTML on ranked URLs + a re-run panel	A reprint of last month’s PDF
DIY hours	Weekly on a calendar	Same panel, same IDs, tickets closed	“Marketing will handle it”
Tools	Monthly or annual seat	Same prompts scored, exportable	A second SEO suite you already pay for

Checklist — the stack is complete only if:

• Each line is named as audit, retain, hours, or tools
• Hours that the agency will not do are still on your calendar
• Tools you already pay for are not added twice
• Schema work is a ticket, not a mystery SKU — see schema markup for answer engines
• Eligibility is a gate, not a hope — see ranked but missing AI Overviews
• No Spurlock package price is sitting in the model because none is published here

If a vendor quote covers retain and you still clock 10 hours/week internally, those 10 hours are line three. They are not “free because we hired.”

How much does an AEO audit cost?

Whatever quote you get for a dated baseline plus a ranked 30/60/90. I will not invent a median audit invoice. Digital Elevator’s public table does not publish a dedicated “AEO audit” SKU. Their freelance/project band of $150–$2,000 describes single-page optimization, FAQ structuring, and basic schema — the wrong object for a visibility baseline (Digital Elevator, July 23, 2026).

Vendor blogs recycle round numbers for “AI visibility audits.” I have not seen a primary survey of those invoices. Treat every audit sticker you cannot re-fetch on a named page as hearsay. Ask for the artifact, then price the artifact.

If the quote is really	You should see	It is not an audit if
A baseline	Frozen 25–40 prompts, dated captures, mention vs citation vs absent vs wrong-fact	One ChatGPT paste from a sales call
Eligibility	Indexing, snippet rules, robots, canonicals, URL Inspection	“We’ll add llms.txt” as the whole technical pass
Entity map	Name, offer, NAP, About vs schema vs footer	A brand workshop with no URLs
Ranked tickets	Five or fewer first-wave URLs with owners	A 40-post calendar
30/60/90	Sequence plus a kill criterion	A retain proposal stapled to the back

Procedure — price an audit without a fake median:

1. Write the artifact list from the table above. That is the SOW.
2. Get two or three quotes against that list, not against “AEO.”
3. Reject any quote that cannot be staffed without a translator.
4. Amortize the one-time number across the quarter you will use it. Do not pretend it is monthly.
5. Do not convert the quote into a Spurlock SKU. There isn’t one on this page.

Red flag in the audit quote	What it usually means	Move
“Implementation included” for half a year	Retain wearing the audit’s name	Split the objects; price retain separately
Citation guarantee	Nobody honest can sign it	Walk
AEO score, no prompt list	You cannot re-run it	Demand IDs
Price far below a month of their own retain	Thin PDF	Pay for diagnosis or don’t; do not call it a program
Schema plugin as the deliverable	Markup without a fact sheet	Read the schema spoke, then re-scope

A paid audit still spends your hours. Those hours sit on the DIY line even when the invoice is “one-time.”

Audit week	Hours you should expect to spend	If you skip them
Access	CMS, Search Console, analytics, GBP if you have one	The baseline is a guess
Fact interview	Offer sentence, NAP, what you will not sell	Models keep quoting the 2023 press mention
Prompt veto	You kill vanity IDs; they do not invent your revenue questions	The panel is their keyword list
Readout	You can restate the 30/60/90 without the PDF	You bought a brochure

I will not invent how many hours that takes at Spurlock. Clock yours. A founder who cannot spare the interview is not buying an audit. They are buying a file.

An audit that skips URL Inspection is a mood board. Google’s own floor is still indexation and snippet eligibility. Paying for “AEO” while nosnippet is on the answer block is how you fund a report about a page the Overview cannot quote.

How much does an AEO retainer cost?

The monthly labor to run a frozen prompt panel, keep entity facts aligned, rewrite page jobs, and recover lost citations. Scope is what belongs in the retain. This section is only the public dollar chatter around that object.

Digital Elevator’s July 23, 2026 table, which they also describe as drawing on pricing research from Eagles Media Enterprises, Words Have Impact, and WebFX — agency marketing, not a government series:

Their row (July 23, 2026 page)	Sticker they published	What the row actually describes
Freelance / gig	$150–$2,000 per project	Single-page work, FAQ, basic schema
Productized packages	$500–$2,500 / month	Template-based AI visibility, limited customization
Mid-market retainers	$2,000–$8,000 / month (lede also cites $2,000–$10,000)	Active monitoring, content, entity work
Enterprise programs	$10,000–$30,000+ / month	Multi-platform campaigns, PR, deep reporting
“AI search benefits” at ~$99 / month	Treat as baseline SEO with AI flavor	Not a purposeful AEO program, in their words

Those figures are theirs, on a marketing page, at one crawl. They move. Cheap retainers often mean recycled SEO with a new acronym. Expensive retainers can mean the same. The tell is whether the SOW names a prompt panel, shipped URLs, and non-goals — not whether the number sits in a pretty band.

You might be in this band if	Planning label (not a quote)	Kill it when
Solid SEO floor; you want a log and light fixes	Their “monitor and maintain” $1,000–$2,500 / month	The month is screenshots only
Competitive category; you will ship pages	Their “active optimization” $3,000–$8,000 / month	Content sprints never touch the five money URLs
Multi-brand, heavy off-site, exec reporting	Their “category leadership” $10,000–$25,000+ / month	You bought PR and called every mention a citation
Invoice under ~$1,500 / month labeled “full AEO”	Their warning band	You needed diagnosis, not a nickname for SEO

Procedure — read a retain quote:

1. Map every line item to panel, entity hygiene, page jobs, or citation recovery.
2. Anything left is classic SEO, design, or reporting. Label it honestly or cut it.
3. Count your remaining hours. Those stay on the DIY line.
4. Compare the monthly number to one incremental closed job or retained account you can name. If you cannot name it, you are buying education. Cap the quarter.
5. Re-fetch any public band you pasted into the model. Stale bands are how last year’s blog becomes this year’s budget.

Ask these before the monthly number matters:

• Prompt IDs are written down and will not rotate for 90 days
• Month one is tickets from the audit, not a reprint of the PDF
• Developer hours (yours or theirs) are named. Schema and nosnippet are not “content”
• Engines in scope are listed. Google-only is a different, cheaper object than ChatGPT + Perplexity + Overviews
• Kill clause: what un-ships the retain at day 90 besides “we didn’t like it”

I will not publish a Spurlock monthly. If you need a number for this studio, that is a visibility audit conversation, not a blog SKU.

What do DIY AEO hours actually cost?

Hours × loaded payroll. Not a marketplace day-rate. Not a studio day-rate. There is no AEO SOC code. If you need a public wage proxy, use it as a label, then throw it out the moment you have your own loaded cost.

The U.S. Bureau of Labor Statistics Occupational Employment and Wage Statistics, May 2025 national estimates (OEWS news-release Table 1):

Occupation (SOC)	What the row is	Median hourly	Mean hourly	Mean annual
Market research analysts and marketing specialists (13-1161)	Closer proxy for the person logging prompts and rewriting pages	$37.87	$43.03	$89,490
Marketing managers (11-2021)	Closer proxy if the “DIY” person is actually a director/founder	$80.19	$85.47	$177,770

Those are wages, not loaded cost. Benefits, payroll tax, unused meetings, and the hour you spent arguing about the prompt list are not in the median. Use finance’s loaded rate if you have one. I will not invent a load factor.

Worked arithmetic — a planning illustration, not a measured program and not a Spurlock envelope:

DIY slice	Hours you might clock	Specialist proxy ($37.87)	Manager proxy ($80.19)
Weekly panel run + log	2 hours / week → ~8 / month	~$303	~$642
Two money-page rewrites	12 hours in the month	~$454	~$962
Schema / fact reconciliation	4 hours in the month	~$151	~$321
Illustration total (wages only)	~24 hours	~$908	~$1,925

Read that table as “clock it,” not “budget this.” If the owner cannot spare the hours, DIY is not cheaper. It is unpaid and unfinished. The 30-day operating gate — freeze a panel, then decide — lives on the DIY spoke. This page only prices the calendar.

Checklist before DIY hours enter the model:

• One name on the calendar, not “marketing”
• Hours are estimated from tickets (panel, About, money URL, schema), not a vibe
• Rate is loaded payroll or a BLS row labeled as proxy, dated May 2025
• No studio day-rate is in the cell
• Tool time is not double-counted as both hours and a second job

Hours that are not AEO, even if someone labels the calendar “AI”:

Clocked as AEO	Usually is	Put it on
Rewriting the homepage because branding felt stale	Brand, not citation	Websites / design
A 40-post outline	Content ops	Cluster work after the five URLs exist
Buying links	PR / SEO	Corroboration only if a fact needs a second source
Tool admin and SSO tickets	IT	Tools line, once
Arguing AEO vs GEO vs AIO	Taxonomy	One house term; stop

OpenAI still splits training crawl (GPTBot) from search surfacing (OAI-SearchBot). robots.txt changes for search can take about 24 hours to apply (OpenAI crawler docs). DIY hours spent “blocking all the AI bots” and then buying AEO is how you pay twice to stay invisible.

What do AEO tools cost?

A seat that re-runs prompts. Not a citation. Vendor stickers below were on the public pricing pages we fetched for this post (September 2026). They go stale. Re-read the URL before the number enters a budget.

Vendor page	Published sticker	What the sticker actually includes	Hedge
Otterly.AI pricing	Lite $29/mo; Standard $189/mo; Premium $489/month; Enterprise from $1,000/month	Prompt counts (15 / 100 / 400); four engines on those tiers (ChatGPT, Google AI Overviews, Perplexity, MS Copilot); Claude / Gemini / AI Mode as add-ons	Annual is 15% off on that page; add-ons are extra
Profound pricing	Starter $99/month; Growth $399/month; Enterprise custom	Starter: ChatGPT only, 50 prompts; Growth: 3 engines, 100 prompts; both billed yearly (“2 months free”) on the page we fetched	The monthly-looking number is an annual commit on self-serve
Semrush AI Visibility Toolkit	Base $99/mo per domain, billed annually on that page	25 custom prompts, 1 domain for brand performance, mentions from ChatGPT, Google AI, Gemini, Perplexity	Extra prompts and domains are add-ons; Semrush One bundles SEO + AI at a higher sticker (Semrush plans)
Semrush One	Starter $199/mo monthly ($165.17/mo billed annually on the page we fetched)	SEO toolkit plus AI visibility (50 prompts on Starter as shown)	Do not add this and a standalone AI toolkit for the same domain without checking overlap

Add-ons are how a “$29” or “$99” sticker becomes a different product. Semrush’s AI Visibility KB lists extra prompt packs at $60/month per 50 prompts and extra brand-performance domains at $99/month per domain (Semrush AI Visibility Toolkit). Otterly’s public add-on table (USD, excluding tax, on the pricing page we fetched) prices extra engines separately — Google AI Mode and Gemini as low as $9/month on Lite, Claude higher — and +$99 per 100 extra prompts on Standard/Premium (Otterly add-ons).

Sticker you repeated in Slack	Cash this month if billed yearly	Why the cell lies
Profound $99/mo	Annual commit on the page we fetched (“billed yearly,” “2 months free”)	The $99 is a monthly-equivalent, not a cancel-anytime seat
Semrush AI $99/mo per domain	Annual on the AI toolkit page we fetched	Plus $60 per extra 50 prompts if 25 IDs are not enough
Otterly $29	$29 if monthly; ~$25 on their annual toggle	Lite is 15 prompts. A 25-ID panel already overflows

Search Console can show generative-AI performance for Google’s own features, and owners can include or exclude the site from those features (Search generative AI control). That control is not a ChatGPT log. Do not pay a tool to replace GSC for Overviews and then skip GSC.

Buy the tool when	Do not buy the tool when
You will re-run the same IDs and export the log	You wanted the dashboard in the sales deck
Two people need the same captures	The only user is the vendor
You already know the five URLs	Eligibility is still on fire
You are comparing engines on a frozen set	You are shopping for a grade

Decision list:

1. If you already pay Semrush for SEO, price the delta for AI visibility — add-on vs One — before you add Otterly and Profound.
2. Prompt count must cover the frozen panel, not a vanity list of 400.
3. Annual billing is cash this month, not “$99.” Profound’s self-serve page is explicit about yearly billing.
4. A GEO audit button inside a $29 plan is still not a human baseline.
5. Tools do not implement schema.org Organization for you. People do.

How do you add the four lines without double-counting?

One sheet, four rows, one total. Hide a row and the founder pays it anyway.

Planning illustration — not a recommended package, not a quote, not Spurlock math. Suppose a team is looking at a mid-market retain in Digital Elevator’s $3,000–$8,000 “active” chatter, already pays Semrush One at the annual Starter equivalent, and will still spend 8 hours/month internally on the specialist proxy:

Line	Planning-band input	What you write in the cell	Do not also add
Audit	One vendor quote for baseline + 30/60/90	That quote ÷ 3 if you amortize across a quarter	A second “discovery fee” for the same artifact
Retain	Their SOW monthly	The invoice	DIY hours the SOW already staffs
DIY hours	8 hours × $37.87	~$303 wages, labeled proxy	Hours the retain already includes
Tools	Semrush One annual Starter equivalent	~$165 / month on the page we fetched	A second AI-visibility seat for the same 25–50 prompts

If the retain already runs the panel, do not book the same 8 hours as DIY and as the invoice. If the retain does not run the panel, the 8 hours stay. Most thin retains hide that swap.

Double-count	How it shows up	Fix
SEO suite + AI add-on + a third tracker	Three invoices, one prompt list	One logger; GSC for Google features
Retain “content” + your editor on the same URL	Two people, one page	RACI: who ships HTML
Audit PDF reprinted as month one of retain	You paid twice for diagnosis	Month one of retain is tickets, not a reprint
Schema plugin subscription + developer hours	Markup billed twice	One owner of JSON-LD
Founder hours billed at $0	The total looks cheap	Put a rate in the cell even if finance hates it

Procedure:

1. List every invoice that has “AI,” “AEO,” “GEO,” or “visibility” in the name.
2. Tag each as audit / retain / hours / tools.
3. Strike duplicates (same prompts, same URLs, same month).
4. Add the remainder.
5. Compare the remainder to one named job or retained account. If you cannot name it, cap the quarter and call it education.

The mix of AEO vs classic SEO inside an organic program is a different spoke (AEO vs SEO budget split). Do not paste that percentage table into this dollar stack and call it a price.

What does a cheap retainer usually drop?

The HTML. The panel becomes a gallery. Entity facts drift. Citation recovery is a screenshot of the loss. You still get a slide.

Digital Elevator’s own warning: low-budget SEO subscriptions that claim “AI search benefits” around $99/month should be treated as baseline SEO with AI-flavored marketing, not a purposeful AEO program. That is their sentence, on their page, dated July 23, 2026. I will not promote it into a law. I will say it matches what I see when the SOW cannot name a URL.

Line item that vanished	What you are left holding	What to demand instead
Frozen prompt IDs	A new list every week so nothing trends	IDs that survive 90 days
Eligibility pass	llms.txt theater	URL Inspection on money pages
Fact sheet	Three About sentences that disagree	One offer sentence in HTML and JSON-LD
Ranked five URLs	A blog calendar	Five tickets with owners
Shipped changelog	“Content in progress”	URLs that went live, dated
Non-goals	“Full-funnel AI domination”	A written will-not-do list

Checklist — the retain is theater when:

• Month-end email is captures with no URL shipped
• Schema was “done” in a plugin and never matches visible copy
• The prompt set changed so the chart could go up
• Google features are the only engine and GSC was never opened
• You cannot read the ranked list out loud in five minutes

If you are going to buy cheap, buy logging (a tool in the Otterly Lite neighborhood) plus DIY hours you will actually calendar. Do not buy a retain that has already dropped the work. A $29 logger plus unpaid hours can still fail — but at least you did not rent a report.

What failure mode burns the budget first?

Paying for inclusion while the page cannot be quoted. Rank is not a citation. Ahrefs’ overlap work and Google’s own eligibility rules live on ranked but missing AI Overviews. The cost version of that failure is simple: the retain keeps invoicing while noindex, nosnippet, or a JS-only answer block keeps you out of the candidate pool.

Second place: schema soup billed as AEO. Markup that does not match the page is a conflict, not a feature. The schema spoke is the fix list. This page only prices the mistake — you paid a line item that trained crawlers to distrust you.

Failure	Cash you still spend	What it costs besides cash	Do this instead
Ineligible money URL	Full retain month	Zero chance at an Overview supporting link	Inspection first; pause retain tickets on that URL
Fact conflict (About vs schema vs footer)	Hours rewriting the wrong sentence	Models quote the easiest lie	Freeze one fact packet
Tool without a panel	Seat	Anecdotes	Freeze IDs, then log
Content blast, no five-page set	Writer retain	Models still quote a vague homepage	Homepage, money URL, About, proof, location/contact
Blanket bot block	Audit + retain	You opted out of the answers you wanted	Split training vs search crawlers
Founder hours at $0	Looks cheap on paper	The program dies in week three	Put hours on the stack

Concrete failure, unnamed client, not a result claim: a site ranks for the commercial query, Overview cites a competitor, the agency sends “AEO report” screenshots, and nobody has cleared data-nosnippet on the hero. You did not underpay. You paid for the wrong object.

• URL Inspection on the five URLs before month two of any retain
• Offer sentence identical in HTML and Organization JSON-LD
• robots.txt does not disallow the search bots you are paying to impress
• Panel IDs frozen before the first invoice after the audit
• One named internal owner even if an agency runs the captures

If those boxes are empty, the cheapest honest move is to stop the retain, run eligibility, and restart the stack. Burning another month because the contract auto-renewed is not thrift.

How should you pick a planning band this quarter?

From the object you are missing, then from LTV you can name — not from a pie chart on someone else’s blog. This is a dated starting range. Recalibrate when the panel and the pipeline disagree.

Situation this quarter	Planning-band shape	What you fund	What you do not fund
Cannot describe current citations or eligibility	One-time audit quote	Baseline + 30/60/90	A year of retain on hope
Baseline exists; nobody ships	Retain against a named URL list	Panel + page jobs	A third tool
Named owner, eligible pages, small category	DIY hours + one logger	Specialist hours + Otterly-class seat	A productized retain that dropped HTML
Competitive category, ranked tickets, real LTV	Mid-market retain chatter ($3k–$8k in Digital Elevator’s “active” row) as a conversation starter	Labor that ships	Citation guarantees
Entity/crawl fire	Eligibility only	Technical + fact sheet	New acronym, new calendar
Low LTV, one extra job cannot cover 90 days of retain	DIY or audit-sized only	Cap	Category-leadership enterprise row

Procedure:

1. Name the missing object (baseline vs shipping vs logging).
2. Fill the four-line sheet with labeled bands and one real quote where you have it.
3. Date the sheet.
4. Set a kill date (end of quarter is enough).
5. Put the cousin decisions on the right pages: worth-it, DIY vs hire, retain vs audit. Do not re-litigate them here.

Do not pick the band from	Pick it from
A recycled “$3k–$15k” roundup with no primary URL	A named SOW plus your hours
Spurlock rumor	Nothing — we did not publish a rate here
Last year’s SEO retain plus 20%	Jobs: panel, facts, pages, recovery
Tool list in a Twitter screenshot	Vendor pages you opened this week
Fear of Overviews	A 20-prompt log you already ran

If leadership demands a single number in the first meeting, give them the four-line total for this quarter’s object, with every public figure tagged “planning band, dated [source date].” That is a decision. A fake industry average is not.

How do you measure whether the spend is working?

On the same frozen prompt IDs you paid to run, plus eligibility, plus conversions on protectable URLs. Not on restored 2022 informational CTR. Not on a vendor AEO grade. Not on a conversion multiplier from a blog.

Scoreboard column	What you log	Working looks like	Not working looks like
Eligibility	Inspection, index, snippet	Money URLs 200, snippet-eligible	You funded AEO on a blocked page
Citation / mention	Same IDs, same wording, engine, date	Mention → citation, or wrong-fact → accurate	New IDs every week so the chart rises
Accuracy	Offer, NAP, price facts in the answer	Hallucinations drop on brand prompts	Models still invent a package
Shipping	Changelog of URLs	Tickets closed against the 30/60/90	Zero HTML, fat PDF
Pipeline	CRM notes that name an AI product	Assisted opportunities you can read	A 4.4× slide

Search Console’s generative-AI report is the Google-features column. ChatGPT and Perplexity need the panel. Do not merge them into one “AI traffic” number and then argue with finance.

• Week-1 and week-12 captures exist for the same IDs
• Each ID is revenue or vanity; vanity cannot defend the retain
• Tool invoice maps to those IDs (prompt count ≥ panel size)
• No unsourced multiplier is in the ROI cell
• DIY hours actually appeared on the timesheet

If citations are flat after you shipped the ranked five, you have a corroboration or extractability problem — not proof that you “need a more expensive retain.” If citations never had a chance because of nosnippet, you have an eligibility problem. Price the fix, not a bigger acronym.

What should you skip if you only have a week?

Skip a tool bake-off, a six-month retain, and a 40-post calendar. A week can price the stack. It cannot earn a citation.

1. Write the four lines on one page. Leave dollars blank if you must. Do not leave the names blank.
2. Run 15–25 prompts you did not invent this morning. Screenshot. That is the seed of a panel, not a program.
3. URL-Inspect the homepage and one money URL. If either fails snippet eligibility, the next dollar is a fix, not a retain.
4. Open the three vendor pages and Digital Elevator. Copy stickers with dates. Label them planning bands.
5. Clock last month’s real hours that were already “AEO.” Put a rate next to them.

Week impulse	Why it wastes the week	Do this instead
Buy three loggers	Overlap, no IDs	One logger or a spreadsheet
Sign an annual Profound-class seat to “look serious”	Cash this month for a panel you do not have	Freeze IDs first
Hire a $500 productized package	Template FAQ on the wrong URLs	Inspect, then decide audit vs DIY
Rebuild schema sitewide	Soup	One Organization node that matches About
Argue AEO vs GEO vs AIO	Acronym tax	One house term; see the playbook

A week is enough to refuse a bad retain. It is not enough to declare ROI. If the week ends and nobody owns next month’s hours, you do not have a cost model. You have a doc.

When is paying for AEO services not the next dollar?

When the missing object is eligibility, a fact packet, or a named owner — not a bigger invoice. Pay for AEO services when you know what you are buying and someone will operate it. Do not pay when the site cannot be quoted, the offer still changes every sprint, or the only KPI is “rank in ChatGPT.”

Gate	Next dollar	Not the next dollar
Money URL fails Inspection	Developer + eligibility	Retain
No frozen prompts	Spreadsheet + 2 hours	Enterprise monitoring
Facts disagree across About / schema / footer	Fact owner for 30 days	Content blast
Nobody will log for 30 days	Wait, or buy an audit you will actually read	A retain that needs that person anyway
LTV cannot cover 90 days of mid-market retain chatter	DIY + cheap logger, or stop	Category-leadership band
You already have a ranked 30/60/90 and no ship capacity	Retain or staffed DIY	A fourth tool

The worth-it spoke is the philosophy of waiting. This page only says: do not spend retain money to avoid a robots.txt conversation. Do not spend tool money to avoid writing an offer sentence. Do not spend DIY hours you will not calendar and then blame the acronym.

I will not invent a break-even citation count. If finance cannot name LTV, you cannot name ROI. Cap the stack at education, or stop.

FAQ

How much do AEO services cost?

A stack: one-time audit, monthly retain, DIY hours, and tools — added, not swapped. Public mid-market retain chatter on Digital Elevator’s July 23, 2026 guide clusters around $2,000–$10,000 per month; productized packages sit lower; enterprise rows sit much higher. Those are planning bands, not Spurlock Studios pricing. Tool stickers and DIY payroll are separate lines on the same sheet.

How do I measure whether much do AEO services cost is working?

Re-run the same frozen prompt IDs and score mention, citation, and accuracy, then pair that with URL Inspection and conversions on protectable URLs. If the panel IDs moved, the delta is fake. If informational CTR fell but citations and brand search held, you are measuring the wrong cousin. Search Console’s generative-AI report covers Google’s features only — it is not a ChatGPT log.

What usually fails first when teams try this?

Eligibility and the scoreboard. Teams buy a retain while the money URL is nosnippet or the prompt list changes every week so the chart can rise. The cheap failure is a tool with no hours. The expensive failure is a retain that emails screenshots and never ships HTML.

How long does this take to show results?

A four-line cost sheet and a seed panel can exist in a week. An audit readout is a project, not a morning. Directional citation movement on some prompts is a quarterly object, not a week-two promise. Restoring pre-Overview informational CTR is not a 90-day guarantee and often not a guarantee at all.

What should I skip if I only have a week?

Skip the tool bake-off, the annual seat, and the productized FAQ package. Write the four lines, inspect two URLs, screenshot 15–25 real prompts, and copy vendor stickers with dates. That is a cost model. A week of new posts is not.

When is this not worth doing yet?

When crawl or snippet eligibility is on fire, entity facts still churn, or nobody will own the hours. Fix the floor, or buy a bounded audit you will read, or wait. A mid-market retain against a site that cannot be quoted is a report subscription. The worth-it gate is the longer argument; the cost version is: do not add line two until line one has an object.

CTA

Stop shopping for one AEO sticker. Add the four lines, date every public band, and refuse a retain that cannot name the HTML it will ship.

Lane: /visibility · Book a visibility audit.