The suspicion usually arrives quietly. The monthly invoice comes. The report comes. The numbers look fine in isolation. But the business has not changed. Leads from organic search are flat. The pages your agency promised to rank are still on page two. Six, nine, twelve months in, the results that were sold to you have not arrived.
Before you act on the suspicion, you need to separate two very different situations. The first is an underperforming agency producing little or nothing for the fee. The second is a competent agency working on a realistic timeline in a competitive niche where four to twelve months of patience is genuinely required. The actions you take are different in each case, and getting the diagnosis wrong is expensive in both directions.
This guide walks through the specific signals that confirm the agency is the problem rather than the timeline. It covers how to measure delivery objectively against the contract, the structured conversation to have with the agency before you decide anything, and what a clean transition looks like if you decide to leave. It is written for the person paying the invoice and accountable for the outcome, not for anyone selling SEO.
Key Takeaways
- SEO genuinely takes four to twelve months for low-competition keywords and longer for competitive ones, so a flat result at month three is not by itself proof of agency failure.
- Three signals together usually confirm the agency is underperforming: deliverables that arrive late, light, or not at all; reports that emphasise vanity metrics over the metrics the contract specified; and an inability to explain in plain language what the strategy is and why it is not working yet.
- Before switching, run a structured review meeting that asks for the deliverable inventory, the strategic rationale, the gap to target, and a written remediation plan with dates, so you can decide on evidence rather than frustration.
Separate underperformance from a realistic timeline
The first job is honesty about what SEO is supposed to deliver inside the period you have been working with the agency. Most public guidance and Google’s own commentary place meaningful ranking gains for low-competition keywords in a four to twelve month window, with high-competition terms taking two years or more. If you are three months in on a competitive keyword set, the absence of movement is not by itself a verdict on the agency.
The honest test runs in two directions. First, what was promised in the proposal and contract and on what timeline. Second, what is the realistic timeline for the keyword difficulty, market and asset base you actually have. If the gap between the two is wide, the agency oversold at the start, which is a real problem but a different problem from current underperformance. Confirming agency failure means confirming that the work being done is insufficient or wrong, not simply that the calendar has not yet matured.
What “results” should mean in the contract
Look at the contract and the original proposal. Results should be defined in three layers. Activity layer: deliverables produced, such as pages published, links earned, technical fixes implemented. Ranking and traffic layer: target keywords and their movement, organic sessions, click-through rate. Business layer: organic-attributed leads, demos, signups or revenue. If the contract only specifies the activity layer, you have no real basis to measure delivery in business terms, and that is itself a finding.
What a fair timeline looks like
For most B2B and SME contexts, the first ninety days are foundational work: audit, fixes, content plan, baseline links. Months three to six should show ranking movement on lower-competition target terms. Months six to twelve should compound into traffic and conversions. If a year has passed and the activity layer is thin and the ranking layer is flat across all targets, the timeline argument no longer protects the agency.
The signals that confirm underperformance
Three categories of signal, taken together, indicate the agency rather than the timeline is the problem. Any one of them in isolation can have a benign explanation. Two or three of them together rarely do.
Deliverable shortfall
Compare the deliverables in the contract or scope of work to what has actually shipped, line by line. Pages promised versus pages live. Links promised versus links acquired and indexed. Technical fixes promised versus fixes verifiable in the source code or in Search Console. A pattern of late, light or absent deliverables is the cleanest signal of underperformance because it is the easiest thing to measure objectively.
Reports that drift toward vanity metrics
If the original contract framed success in terms of organic leads or conversions and the monthly report has gradually shifted to impressions, total keywords ranking, and average position across the entire domain, the agency is reframing success because the original frame is not flattering. Look for whether the report still answers the original question: is the work generating the business outcome we agreed on. If the report has stopped answering that question, that is a finding.
No coherent strategic explanation
Ask the account manager, in plain language, what the strategy is for the next ninety days and why it is the right strategy for your situation. A competent agency can answer this with reference to your keywords, competitors, content gaps and link profile. An underperforming agency tends to answer in generic terms or by listing activities rather than connecting them to outcomes. If you cannot get a clear strategic narrative, the absence of one is itself the answer.
How to measure delivery objectively
Before any difficult conversation, build a simple, defensible measurement of what has been delivered. The point is not to win an argument but to have facts you can stand behind whatever you decide.
The deliverable inventory
Pull the contract, list every deliverable with its committed cadence and quality bar, and mark each one as delivered, partial, or missing for the period under review. For content, count published pages and check word count and topical relevance against the brief. For links, request the source list and check that the links are live, dofollow where promised, and on relevant sites. For technical work, request the change log and verify it against your CMS or staging environment.
Ranking and traffic gain on target keywords
Pull rankings for the agreed target keyword list at start of engagement and now. Movement is the metric, not absolute position. Pull organic sessions and organic-source conversions from analytics for the same period, with a comparable baseline. If the report your agency sends does not allow you to do this independently, ask for raw data export, dashboard credentials or a working session to walk through it.
Business outcome attribution
The hardest layer, and the one where many engagements quietly fail. Match organic-attributed leads or revenue against the engagement period. If the agency cannot show this, request that they help you stand it up before any further fee is paid. The inability to attribute outcomes is itself a finding when the contract specified them.
The structured conversation before you decide
Once the inventory is built, schedule one meeting with the senior person on the account, not the account manager, and run it as a review rather than a confrontation. The objective is to give the agency a fair, documented chance to explain and remediate before you decide.
What to ask for in the meeting
Walk through the deliverable inventory and ask for explanation of any shortfalls. Ask for the current strategy in plain language, including which keywords are the priority, why, and what is planned for the next ninety days. Ask for the gap-to-target analysis: where the rankings need to be by when, and what specifically will move them. Ask for a written remediation plan with dated milestones, and ask whether any portion of the recent fee should be credited given the shortfall.
What good and bad responses look like
A good response acknowledges where delivery has fallen short, attributes it to specific causes, and offers a concrete corrected plan with dates. A bad response defends the work in general terms, deflects to algorithm changes or your industry being difficult, and proposes no measurable change. The quality of the response is information.
What happens if you decide to switch
If you decide to leave, the transition matters more than the decision itself. A clumsy transition can erase the small amount of value the original engagement did create.
Take ownership before you give notice
Before any termination notice, make sure you own and control the assets. Domain registrar access, hosting, Google Search Console, Google Analytics, Google Business Profile, the CMS, any link-tracking tools and any content drafts. Confirm in writing who holds what. The most common transition failure is discovering that an account or asset is in the agency’s name and recovery becomes a months-long negotiation.
Diagnostic first, strategy second
Resist the urge to sign the next agency to a long contract on the strength of a sales pitch. The first month with a new provider should be a diagnostic, not a strategy. They should audit what was actually done, what is salvageable, what needs to be undone, and what the priority targets should be. Only then should they propose the engagement shape and price.
Expect a settling period
Switching agencies almost always means a quiet quarter while the new team takes over instrumentation, finishes diagnostic work and sequences priorities. Plan for this in your forecasts. The compounding starts again from the new strategy and is rarely visible inside ninety days.
Conclusion
The decision to leave an underperforming SEO agency is rarely made on a single piece of evidence. It is made when the deliverable inventory shows shortfall, the reports have drifted from the metrics that matter, and the strategic explanation does not survive a direct question. When all three are true, the timeline argument no longer protects the engagement.
The work to do before that decision is more useful than the decision itself. Building the inventory forces clarity about what was actually promised. Running the structured review gives the agency a fair chance and gives you documented evidence either way. Securing the assets before notice protects the value that does exist. Whatever you decide afterwards rests on better ground.
Frequently Asked Questions
How long should I wait before deciding my SEO agency is not delivering?
Are flat rankings always the agency’s fault?
Should I withhold payment if I think the agency is underperforming?
What single metric most reliably indicates real delivery?
How do I find a replacement agency without repeating the same mistake?
Can I salvage the work the previous agency did?
If you are reviewing whether your current SEO engagement is delivering and want a structured second opinion before you decide, you can enquire now.
