Why a free trial isn't a trial
Almost every agency will offer you something free before you sign — an audit, a month of "strategy", a competitor teardown. Take it if you want. Just be clear that free work is produced by the side of the agency that sells, on a template, and it ends in a PDF. We've written at length about why the free audit in your inbox is a sales asset.
A pilot is different in kind, not in size. It isn't a smaller pitch; it's a smaller job. The thing you're trying to find out is whether this agency can get a change through your CMS, your developer's sprint and your approval chain, and whether they describe accurately what happened when they tried. None of that is observable in a document.
Billing is part of the instrument. Work with an invoice attached gets staffed with someone whose hour has a number on it. Work with no invoice gets whoever is free on Thursday. You are specifically trying to watch the first kind of work happen.
The three things a 60-day pilot must actually ship
A fair pilot ends with evidence you can open on the last day. Three artefacts, and not one of them is a slide.
- A frozen baseline. The number the eventual retainer will be judged against, captured before anything changes and signed by both sides. Qualified leads from organic search over the trailing 90 days is the honest metric; sessions is the easy one. Write the definition down with worked examples — what counts as a qualified lead is an argument worth having in week one rather than month seven.
- One technical fix merged into production. Not specified, not ticketed, not "ready for release". Merged, live, and verified on the live URL. Pick something real: a rendering fault, a canonical mess, an indexation block on a template that matters. You're watching how they handle your developer, your release train, and the fortnight when nothing deploys because of a launch.
- One content set published. A brief, a draft, an edit, and live URLs. Three to five pages is plenty. You learn more from watching one brief survive your legal review and your founder's opinions than from any content strategy.
The four things a pilot must never contain
These are the terms that turn a test into a trap. Each one is common enough that you should read for it specifically.
- A promised ranking position. Google says plainly that no one can guarantee a #1 ranking, and warns about agencies that claim a special relationship with it. An agency promising a position inside 60 days is either buying links or planning to argue about the wording in week nine.
- A projected traffic number. A single percentage in a pilot proposal is a sales instrument, not a forecast. A range with its assumptions written next to it is genuinely useful. "+180% organic traffic" is neither.
- An auto-renew. The pilot has to end. If it converts, it converts because you signed something new on a date you chose. A clause that rolls into a retainer unless you cancel is a subscription trap wearing a trial's clothes.
- Anything you can't take with you. Content in source form with full rights, crawl exports, the baseline sheet, the link prospect list. All of it yours on day 60 whatever you decide.
Fixed price, fixed end date, and what a pilot should cost
Two numbers and one date, agreed before anything starts. Open-ended pilots fail for a boring reason: with no end date nobody has to decide anything, and week nine arrives with the same conversation you were having in week two — except you've now paid for nine weeks and lost your comparison set.
Price it against the retainer it's testing. A pilot is the same work at a smaller scale, so it costs roughly one to two months of that retainer. There's no volume discount on senior time, and a pilot priced far below the retainer is being subsidised by the retainer they expect you to sign — a discount you repay later with interest.
- Testing a ₹40,000/mo retainer: a fair pilot is roughly ₹40,000–₹75,000. A baseline, one fix, three pages.
- Testing a ₹75,000/mo retainer: roughly ₹75,000–₹1,50,000. A baseline, one or two fixes, four or five pages.
- Testing ₹1,50,000/mo and above: budget a full month at minimum. On a large site the diagnosis alone eats the first three weeks, and a pilot that skips it is guessing.
| Fix in writing | Why it matters | What happens if you don't |
|---|---|---|
| The fee, ex-GST, flat | A pilot with an hourly meter is not a fixed scope | The overrun arrives in week seven, when you no longer have other quotes |
| The end date, counted from access | Sixty days from full access, not from signature | Your access delays quietly become their extension |
| The three deliverables, named | "A pilot project" is not a scope | You get a deck, and a disagreement about whether it counted |
| Who does the work, by name | A pilot is the most seniority-dependent thing an agency sells | The pitch team runs the pilot and a junior inherits the retainer |
| What you keep on day 60 | Everything produced, in source form, no exceptions | The baseline sheet stays with them and cannot be reused |
| That it does not auto-renew | Renewal should require a signature from you | Month three turns up as an invoice |
How the output is judged, and who does the judging
Sixty days is too short to judge Google and exactly long enough to judge an agency. Almost every pilot that goes wrong went wrong by confusing the two.
The arithmetic isn't on your side. Google's own documentation says indexing is never instant — a new page can take about a week to be crawled and indexed, and validating a fix in Search Console typically takes up to about two weeks. Add an approval cycle at your end and a page published in week five has barely been looked at when the pilot closes. Judging it on its ranking is judging a queue.
- Did the three artefacts ship? Binary. "In review" is a no. A pilot ending with two of three shipped is a useful result rather than a failure — provided the written note says which one stalled and where.
- Was the diagnosis right? Ask in week one for the single biggest thing capping the account, in one sentence. On day 60, check whether the evidence agreed. This separates an agency that read your site from one that read your industry.
- Impressions for the queries in the plan. Not sessions, and not average position across every query you accidentally rank for. Impressions move first, often inside four to six weeks, and Search Console reports them per query so the claim is checkable.
- Did the written note contain something you didn't want to hear? Every real engagement produces one in the first two months. A pilot report with no bad news is a pilot report nobody wrote.
- Who judges: the person who will own the retainer internally, plus the developer who took the fix. Not the founder alone — the founder wasn't in the sprint and will grade on how the calls felt.
How a pilot converts into a retainer, or ends without penalty
Conversion terms belong in the pilot agreement, not in a negotiation on day 61 when one side has all the information and the other has a sunk cost.
- The retainer price, fixed on day one. If the monthly number can move after the pilot, the pilot's findings can move with it. Anything that would trigger a re-quote should be named up front.
- A credit, or an honest reason there isn't one. Crediting the pilot fee against the first retainer month is common and fair. It also creates a mild incentive to recommend a retainer, and an agency that says so unprompted is telling you something about how it thinks.
- No penalty for stopping. No exit fee, no clawback, no "discovery costs" invoice. You bought 60 days and you paid for 60 days.
- Handover inside five working days, as working files rather than PDFs.
- A written recommendation, including the option to say no. The most valuable pilot outcome we've seen is "your constraint is the pricing page, not organic search — spend this elsewhere this quarter." An agency that can't write that sentence isn't running a trial.
What we sell instead of a pilot, and why
We don't sell a 60-day pilot, and we could — it packages easily and it converts. The honest reason we don't is that the first quarter of our retainer already is one, with the risk pointed at us instead of you.
On day one we freeze your trailing-90-day count of qualified leads from organic search. Both sides sign the number and the definition of "qualified". If we haven't beaten it in 90 days, we keep working — same scope, same team, no invoice — until we do. We never promise a specific ranking position, because nobody controls Google's index. A pilot asks you to pay to find out whether we're any good; this arrangement asks us to pay if we aren't.
SEO is ₹75,000 a month ex-GST, or ₹40,000 for sites under roughly 50 pages. The first quarter is committed, because 90 days is the shortest window in which SEO can be judged fairly and it's the window the guarantee runs on. After that it's month to month with 30 days' notice and you keep every asset. We take three new clients a month, because free months are real months and a promise you can't survive at volume isn't a promise.
If you want something smaller first, the fixed-scope audit is ₹40,000 in ten working days, credited in full if you start a retainer within 60 days. We'll label it accurately: it ends in a document, which by the standard on this page makes it a diagnosis rather than a pilot. And if you're collecting comparable quotes from several agencies, writing the brief properly does more for the comparison than any trial will.