Why Buying SEO Feels Like Buying a Mystery Box
Done-for-you SEO means handing the entire operation to an external provider. Technical audits, on-page optimization, content creation, link building, reporting. You run the business. Someone else handles organic traffic.
The pitch is seductive, and I understand why. No internal hires. No learning curve. No guessing what Google means by “helpful content.” Just rankings, traffic, and leads, delivered on autopilot.
Here’s the problem nobody talks about: the market has no trial mechanism. You cannot comparison-shop providers before buying. You cannot evaluate them cheaply while buying. And the 6 to 12 month feedback loop means you cannot evaluate them quickly after buying. (For context on those timelines, see Ahrefs’ study on how long SEO takes)
So the only way to judge a provider is to sign a retainer and pay for months. Providers know this. Some exploit it.
Budget-tier done-for-you SEO runs $500 to $1,500 per month. Mid-tier runs $1,500 to $5,000. Premium engagements start at $5,000. Those figures come from the breakdown of SEO costs for small businesses, which also details what each tier should include. Do the math on a mid-tier mistake: $5,000 to $15,000 gone before you can decide it was a mistake.
This article is about how to buy done-for-you SEO without betting two quarters of fees on a provider you cannot yet judge. You will get a way to manufacture a trial before you sign, the questions that separate real providers from artifact sellers, and a 90-day framework for judging any provider with your own Search Console data.
The Real Risk Isn’t Bad Agencies. It’s the Broken Feedback Loop.
Most advice about done-for-you SEO warns you about “cookie-cutter campaigns” and “quality variability.” That advice is true. It also misses the point.
Bad providers do not survive because they are clever. They survive because the market gives you no cheap way to catch them. Every other marketing channel you buy gives you fast feedback. Paid ads may not have a free trial, but within days you know whether a campaign works, because you see spend, clicks, and conversions in near real time. With done-for-you SEO, the equivalent signal takes months. Your first two quarters of fees are effectively an unsecured loan to the provider.
Meanwhile, the provider keeps shipping artifacts that look like progress but you can’t see if they are meaningful outcomes or just filler for the retainer. PDF audits. Strategy decks. Blog post counts. Keyword spreadsheets. A 40-page audit feels like work. It is a document. Outcomes are impressions, clicks, and query coverage in Google Search Console, those take months to move and they are not always meaningful.
This artifact-versus-outcome gap is what makes SEO so prone to having black sheep. You can receive every promised deliverable on time and still have flat traffic, because deliverables were never the same thing as results.
Contracts make it worse. Some providers require 6, 9, or 12 month commitments and enforce them. One provider in the market charges at least $100 per remaining month to exit early. A long lock-in with no pilot and no exit process is a documented red flag, not a standard term you should absorb.
The practical conclusion: you cannot fix the market, but you can fix your own buying process. The rest of this article gives you two tools. First, a way to manufacture a trial before you sign the contract. Second, a falsification framework that lets you judge any provider within 90 days, using data you already own.
What Done-For-You SEO Should Actually Include
“Full service” is the most abused phrase in this market. Two providers can use the same label while delivering completely different work. So before you compare prices, you need to know what a credible scope actually looks like.
A real scope answers what happens before, during, and after publication, in four areas. First, research and prioritization: business and competitor discovery, keyword research tied to search intent, and a prioritized roadmap with a stated reason for each page. Second, production and quality control: who drafts, edits, fact-checks, and approves every page. Third, technical and publishing work: who publishes in your CMS, what the audit covers, who implements fixes, and whether there is a rollback path for unsafe changes. Fourth, measurement and maintenance: Search Console and analytics access, reporting on impressions, clicks, and conversions, and a refresh process for pages gaining or losing traction.
If a proposal is silent on any of these four, that silence is what you’re going to buy.
“Full service” also has boundaries, and providers rarely volunteer where they sit. Ask before signing whether these are included: link building and digital PR outreach, technical SEO implementation such as site speed work and structured data, content refreshes and pruning of weak pages, original content creation and recovery work after a traffic drop or penalty. If an item matters to you, require it in the statement of work. Never assume the phrase “full service” covers it.
Ownership is the other fault line. Retain ownership or durable access to your domain, CMS, Search Console property, analytics, content, and reporting history. Grant providers minimum access, and confirm before signing what happens on cancellation: whether content stays, whether drafts export, and whether the provider removes its users. Any process that makes your organic history hostage to one vendor is a red flag.
One decision protects you more than any contract clause: you administer the Search Console and analytics accounts yourself and grant the provider access, not the other way around. This preserves your ability to evaluate them independently, with data they cannot polish, and it sets up the 90-day framework later in this article.
How to Manufacture a Trial Before You Sign
The market will not give you a trial. So build one.
The core move is simple: turn every promise into a concrete responsibility, output, cadence, and owner before money changes hands. A provider’s pitch is not the deliverable. The written commitment is.
Start with a bounded pilot. Define a small page set, the expected outputs, an evaluation date, and what results would justify expansion. A pilot with a written evaluation date converts a blind purchase into a testable bet. Without that date, a pilot is just a smaller blind purchase.
Next, write leading indicators into the statement of work. Do not accept “we will improve your SEO.” Accept this instead: “impressions for the target query cluster will be up versus the frozen baseline by day 90, measured in Google Search Console, or we explain why in writing.” The difference matters. One is a mood. The other is a falsifiable claim you can check.
Before any work ships, freeze a pre-change baseline. Export your Search Console performance data: impressions, clicks, average position, and your query list. Record the date. Without a frozen baseline, every later report is unverifiable, because neither you nor the provider can prove what changed.
Comparison shopping needs identical inputs. So send the same one-page scope to every provider you evaluate: your business goal, priority audience, current bottleneck, pages in scope, approval owner, required integrations, risk constraints, and the metric you will review after the pilot. Identical scopes make proposals comparable. Different scopes make proposals noise.
Negotiate exit terms up front, not when you are unhappy. Contract length. Cancellation terms. What requires an additional fee. Whether there is a pilot or export rights. Get this in writing before the first invoice. A provider who refuses any exit mechanism is telling you something, and you should believe them.
Finally, ask for the actual workflow, not the label. Who does each step. What is automated. Where human judgment enters. What needs your approval. Two vendors using the same “fully managed” label may deliver completely different levels of strategy, production, and review, and the only way to tell them apart is to walk the workflow step by step.
The Questions That Separate Real Providers From Artifact Sellers
A pilot and a frozen baseline give you the machinery. Now you need the interrogation script. These questions work because each one forces a provider to commit to something checkable, and artifact sellers hate that.
Start with measurement. “How do you measure success, and where does the data live?” Google’s own guidance on hiring an SEO recommends asking exactly this, plus whether the provider follows Google Search Essentials. The correct answer names your Search Console and your analytics. A proprietary dashboard you cannot inspect is not an answer. It is a stage prop.
Then ask the falsification question: “What should be true in my Search Console data at day 30, 60, and 90 if this is working?” A provider who cannot answer has never thought in falsifiable terms. That alone is disqualifying, because it means they have never defined what failure looks like in their own engagements.
Links deserve their own question. “Where do the links come from?” If the answer is vague, something like “we have a network of sites,” or the provider cannot show specific examples, walk away. Low-quality links from private blog networks can trigger manual penalties that take months of work to recover from. Google’s spam policies explicitly call out excessive link exchanges and scaled, unoriginal content.
Content production is next. “Who writes the content, who fact-checks it, and who approves it?” Large publishing volume without a stated review process is a red flag, not a bargain. Volume without quality control is how sites accumulate pages Google ignores.
Now the question providers hope you skip: “What is not included?” The boundaries listed earlier are frequently excluded from “full service.” Get the exclusions in writing before the first invoice, not after the first disappointment.
And ask about the exit before you are desperate for one. “What happens if we cancel?” You keep your content, your data, and your account access. Always, regardless of how good the work looks.
Two more tests complete the screen. First, guaranteed rankings, traffic, or fixed result dates are a walk-away signal. No provider can buy inclusion in organic results or guarantee a ranking. A guarantee is either dishonest or so heavily conditioned it is meaningless. Second, ask for representative work similar in intent and complexity to your site, not just a traffic chart. Whether they have clients in your industry, or sites with comparable domain authority, tells you whether their playbook transfers.
The 90-Day Falsification Framework: Judge Any Provider With Your Own Data
The interrogation script tells you what to ask. This framework tells you what to check. It runs on your own data, and the provider cannot argue with it.
The core move is a shift in the question. Instead of asking “is my SEO working?”, which is unanswerable for months, ask “what should already be false if this engagement is failing?” You can answer that in 90 days using Google Search Console. It is free, and it shows ground-truth data from Google itself.
Why Search Console and not Ahrefs or Semrush? Because Search Console reports actual impressions and clicks from Google, not estimated third-party data. It tracks impressions, clicks, average position, and click-through rate, and it surfaces indexing issues. It is the one data source your provider cannot polish.
Do the setup before day one. You own the Search Console property, the provider gets access, and you export the baseline yourself: impressions, clicks, average position, and the total count of queries your site appears for. Freeze the date. That export is your yardstick for everything that follows.
Day 30: The Operational Check
Day 30 does not test rankings. It tests whether a system exists.
By weeks one and two, a real engagement should have produced access, a baseline, technical risk identification, audience and offer context, approval rules, and a prioritized roadmap. By weeks three through six, you should see first repairs or pages shipped, internal links added, measurement verified, and early crawl and index feedback reviewed.
Here’s the failure pattern: a strategy document, but no shipped work, no verified measurement, and no roadmap with a stated reason per page. That is artifacts, not progress. You have seen this movie before.
Be fair about expectations. Impressions and clicks will probably not have moved yet, and that is fine. Rankings may move sooner or much later depending on the site and the query.
One concrete early signal you can check yourself: request indexing for newly published pages and confirm they appear in Search Console’s pages report. If the service publishes at volume, track what percentage of published content gets indexed within 30 days. A high index rate indicates a technically healthy site that Google trusts. A low index rate signals content quality issues, crawl budget problems, or technical barriers.
Day 60: The Visibility Check
Day 60 is where visibility data starts to matter. Total impressions should be flat-to-rising versus the frozen baseline, even if clicks have not followed. An upward trend in impressions before clicks arrive means Google is indexing the content and showing it for relevant queries.
Query coverage should be expanding too. The count of distinct queries your site appears for should be growing, because each new or improved page adds queries. Flat query coverage after two months of “work” means Google is not responding to anything the provider did.
Average position for your target query cluster should be trending downward as a number, even if the site still sits on page two or three. Position and impression share move earliest. Click data follows.
Watch for the high-impression, low-click pattern. Many impressions with few clicks means the site appears at the bottom of search results pages. That is actually a leading indicator that the strategy is working, and a map of which pages to push higher.
If impressions, query coverage, and average position are all flat at day 60, the correct move is a written conversation, not a cancellation yet. Ask the provider to explain the flat numbers against the leading indicators in the statement of work, and require a dated correction plan.
Day 90: The Verdict
By weeks seven through twelve, a real engagement shows consistent delivery, query-level learning, refresh candidates, decisions on weak pages, and a forecast for the next cycle. Now you apply the falsification test.
If impressions, query coverage, and average position are all still flat at day 90, and the provider cannot point to a specific, dated, verifiable cause, such as a Google algorithm update, a site migration, or a technical blocker they documented at the time, the engagement is not working. Two quarters of fees is the maximum a founder should spend on an unfalsified bet.
One honest caveat. SEO timelines are real. It generally takes roughly 6 to 12 months to see meaningful improvements, and most campaigns show meaningful traffic gains between months four and eight. Flat at day 90 is not proof of failure. It is proof that the provider must now carry the burden of evidence, in writing, with dates.
What flat numbers do mean: the provider loses the benefit of the doubt. The next report must connect activity to search visibility with baselines and dates, or you exit.
Judge the first cycle on whether the provider built a sound system and honest leading indicators. Judge later cycles on qualified organic outcomes. A timeline is useful for accountability, but it is not a ranking guarantee.
How to Read a Provider’s Monthly Report (and Catch the Tricks)
The 90-day framework catches flat numbers. The monthly report is where providers hide them.
Start with the simplest test. An honest report needs only three sentences: here is the organic traffic when we started, here it is now, here is what we did and will do next. If a 40-slide deck cannot answer them, the deck is the deliverable, not the report.
Next, check where the traffic lands, not just how much there is. In one practitioner’s audit, a site pulled a few hundred organic visits per month and looked healthy, until the page-level split showed roughly 90% going to a single blog post that was never designed to convert. Total traffic can double while your pipeline stays empty. Always ask for the page-level split, and be equally skeptical of conversions that exist only in the provider’s dashboard. Confidence without a data source is a tell.
Then check the windows. Compare equal, complete date ranges, and log anything else that changed at the same time: a redesign, a paid campaign, a season. Without that log, you will credit the provider for a redesign lift or blame them for a seasonal dip.
Finally, split brand from non-brand queries in Search Console. If most clicks come from people typing your company name, existing awareness is doing the work, not the provider. Discovery queries from strangers are what you are paying for.
Is Done-For-You SEO Right for You?
Whether done-for-you is the right model for you at all, versus DIY or a hybrid setup, is its own question with its own answer. If you are still deciding, read whether SEO is worth it for small business first. If you are weighing a tool-based workflow against an agency, my SearchAtlas alternative page breaks down what that looks like without a retainer.
One caveat before you sign anything. Done-for-you does not mean you disappear. If you will not answer questions or approve work, no provider can succeed, and you will blame them for it.
The Structural Fix: Judge the Thinking Before Anyone Invoices You
The 90-day framework protects you. It still costs you 90 days and three months of fees, because you are judging the work after the money leaves. That is the ceiling of any buying-process fix.
The deeper fix is to reverse the order: see the strategy and a finished piece of work first, decide whether to pay second. No retainer offers that, because a retainer is priced on the month, not on the work, and a busier-looking month is what it sells.
That reversal is what I built Rank-Hub for. Give the AI consultant your domain and it comes back with your strategy, with the reasoning behind each bet, your topic map, one finished Blueprint, a step-by-step plan with the evidence and the reasoning attached, and one finished article for your site. Every recommendation names the objective it serves, so you can read the thinking the way you would read a provider’s proposal, except this one already did the work.
Apply the same test this article gave you for agencies. Is there a stated reason per page? Does the plan say what it expects to move, in real numbers? Can you check it against your own Search Console data? If the thinking does not hold up, you have lost one sitting, not two quarters.
No card, I pick up the compute bill. Start your first session with the AI consultant.