The best SEO outcomes we've seen, and why they won't happen to you

The best outcomes had almost nothing to do with clever SEO

Every SEO has a favourite story. A site that went from nothing to owning a category. A page still bringing in leads six years after it was written. They're real, and they get retold at conferences as though the tactics caused them.

Look at enough of them and the pattern is uncomfortable for everyone selling SEO, us included. The outsized wins are mostly explained by three things, none of which an agency controls.

Category timing. The demand curve was going up while the work was being done. Somebody covering compliance software the year the rules changed, or online education in 2020, looked like a genius doing competent work. The same work in a flat category looks ordinary, because the arithmetic underneath isn't moving.

Product fit. Search sends people who already want the thing. If it's good and priced properly, they convert and tell others. If it isn't, you've built a very efficient machine for delivering strangers to a page that disappoints them. SEO makes a good product visible and a bad product visibly bad, faster.

Being first to an uncovered query set. A window opens when a category's demand exists but nobody has written the definitive pages. Whoever gets there first accumulates links and age while everyone else is deciding. Those windows close within a couple of years, and they don't reopen for whoever arrives late with a bigger budget.

Tactics matter — bad execution wastes all three advantages. But tactics have never been the reason for an outlier; they're the reason it wasn't squandered.

How a case study actually gets made

Nobody is lying. That's what makes case studies so effective and so misleading.

Here's the manufacturing process. An agency runs, say, forty accounts over three years. Some do well, most do fine, a few go nowhere. At the end someone goes looking for the best three, and those three get written up with charts. The other thirty-seven never appear anywhere.

You read three case studies and infer a hit rate of 100%. The real hit rate was never disclosed, and it's the only number that would have told you anything.

The individual numbers get help too, and none of it requires dishonesty:

• Percentages from a tiny base. Going from 40 organic sessions to 400 is +900%. It's also 400 sessions.
• Brand traffic counted as organic. If the client ran a large campaign in the same period, their brand searches spiked. That lands in organic and looks like SEO.
• Traffic reported instead of leads. Sessions are easy to grow with queries no buyer types. Leads are not.
• A convenient start date. Begin the chart at a seasonal trough or a post-penalty low and recovery reads as growth.
• No cost, no timeline. "3x organic revenue" with no mention of the ₹4,00,000 monthly retainer or that it took 26 months.
• The site that was already strong. A twelve-year-old domain with 4,000 referring domains isn't a comparable starting point to yours.

The curve that is actually repeatable

Strip out the outliers and what's left is a shape. It's the same shape on most sites with a real category, competent work and no penalty history, and it's dull enough that nobody puts it on a slide.

The horizontal axis matters more than the vertical one — which is why we ask for eighteen months of intent and give you a 90-day proof point.

• The good outcome at month 18 is usually "organic is now our cheapest reliable lead source", not a multiple.
• The great outcome is that plus a handful of pages that keep earning with no further spend.
• The bad outcome shows up early and looks specific: flat impressions at month three. That's the warning sign worth acting on.
• The worst outcome is eighteen months of traffic growth and no change in leads. It happens, and it means the queries were wrong from month one.

Why the first four months look like nothing is happening

This is the part that gets agencies fired, so it's worth explaining mechanically rather than asking for patience.

A new page has to be discovered, crawled and indexed before it can rank at all — days to weeks, depending on how much crawl attention your site gets. Then it enters the results somewhere unflattering, because Google has no click data on it and no links pointing at it. It sits there while it accumulates both.

Links move slower still. One earned in month two has to be found, evaluated and factored in, and its full effect on a competitive query can take another quarter. Meanwhile your month-two pages are juveniles competing against pages that have ranked for four years.

So months 0–4 aren't idle. They're the part of a fixed deposit where you can see the balance but not the interest. The mistake is measuring them with the wrong instrument — you look at revenue, see nothing, and conclude nothing is working. Look at impressions for your target queries and indexed-versus-published instead. Those move first, and if they don't move by month three, something is genuinely wrong. What actually happens in the first 90 days breaks it down week by week.

There's a client-side half nobody warns you about, too. The most common cause of a slow first six months isn't the agency — it's a technical fix sitting in a dev queue for eleven weeks, or three articles waiting on an approval nobody owns.

What eighteen good months look like in rupees

Numbers make this concrete, so here's a worked model. Illustrative arithmetic with made-up-but-plausible inputs — not a client, not a result, and you should run it with your own figures.

Take a B2B services company on ₹75,000 a month. Eighteen months costs ₹13,50,000 plus GST, plus the internal time, which is real and usually undercounted — call it a day a week from someone senior.

Say the work produces 90 pages, and by month eighteen the site is getting 4,000 non-brand organic sessions a month. At a 2% enquiry rate that's 80 enquiries, maybe half of them qualified, and at a 25% close rate that's 10 new customers a month. If a customer is worth ₹1,00,000 in year one, the channel is paying for itself several times over.

Change one input and it inverts. If the customer is worth ₹8,000, the same effort is a rounding error and the money belonged in ads. Which is why we ask about deal value before keywords, and why the honest answer to "how much can SEO grow my business" is a model you run before you sign, not a number quoted at you.

What our 90-day guarantee actually measures

Given everything above — flat first quarter, eighteen-month curve — a 90-day guarantee sounds like a contradiction. It isn't, because of what it measures.

On day one we freeze your trailing-90-day count of qualified leads from organic search. That's your baseline: your own number, from your own CRM, agreed before any work starts. Ninety days later we compare. If we haven't beaten it, we keep working free until we do.

Note what that is and isn't. It commits us to movement against your own starting point, which is achievable in 90 days on almost any site that isn't already well optimised — the first quarter usually contains recoverable losses, unindexed pages and pages ranking 8–20 that just need finishing. It is not a promise that the eighteen-month curve arrives in three months.

It's also not a ranking promise. Nobody controls Google's index, and any agency guaranteeing a specific position is either buying links or counting on you not to check. Why ranking guarantees are a lie is the full argument.

The reason we can carry the risk is that we take three clients a month. That constraint isn't marketing — it's the only way the maths works when you might owe someone free months.

The outcomes we refuse to promise

Short list. Every item on it is something we've been asked for and said no to, and every one of them is routinely promised by somebody else.

• A specific keyword at position one. Not ours to give. The index isn't a vending machine.
• A revenue multiple. Your ceiling is set by your category's search demand and your deal value, both of which existed before we arrived.
• Results in 30 days. On a site with no history, 30 days is barely enough for indexing. Anything moving that fast was already there, or was bought.
• Recovering a core-update loss by a set date. We can diagnose it and fix the causes. Nobody can date the recovery — it depends on a re-evaluation cycle we don't control.
• Growth in a category with no demand. We'll run the model with you and tell you to spend on ads instead. It's a worse quarter for us and a better year for you.
• That we'll be cheaper than the last agency. We're often not. We're testable, which is a different and more useful property.

How to read someone else's case study without being fooled

Six questions. They take four minutes and they'll disqualify most of what you're shown — including, on a bad day, ours.

1. What was the absolute starting number, not the percentage? 40 to 400 sessions is +900% and means nothing.
2. Was brand traffic excluded? If not, you may be reading the results of somebody else's ad campaign.
3. Leads or traffic? If the headline metric is sessions, ask what happened to enquiries. If they don't know, nobody was measuring the thing that pays.
4. How long, and how much? A result with no timeline and no retainer figure isn't a result, it's a poster.
5. How old was the domain and how many referring domains did it start with? A strong site rebuilding is a different exercise from a new site starting.
6. How many other clients did they take on in that period? The hit rate is the only honest metric in the whole genre, and it's the one that never appears on the page.