The revenue point where an agency starts making sense
There's a revenue level below which hiring an SEO agency is usually premature. It isn't a secret number and it isn't a sales filter — you can work it out backwards from the invoice in about two minutes.
Serious SEO in India starts around ₹75,000 a month, with smaller and simpler sites viable from ₹40,000. Call that ₹4.8–₹9 lakh a year, and assume twelve months rather than three, because compounding is the entire point of the channel. Now apply the one budgeting rule that survives contact with reality: while a channel is still unproven, it shouldn't eat more than roughly 5% of revenue. ₹9 lakh at 5% is ₹1.8 crore. Round it up and you get ₹2 crore in annual revenue.
That's a rule of thumb, not a law. It flexes on three things: what a customer is worth, whether your buyers actually search, and whether your site can take an enquiry when they arrive. Get those three right and the threshold drops. Get them wrong and ₹20 crore of revenue won't save the engagement.
| Annual revenue | What usually makes sense | The catch |
|---|---|---|
| Under ₹50 lakh | Do it yourself, plus a one-off audit once the site is built. | A retainer at this size is 15–20% of revenue on one channel. That's not a bet, it's a hostage situation. |
| ₹50 lakh – ₹2 crore | A freelancer, or a scoped project with a defined deliverable. | You're buying a specific fix, not continuous capacity. Judge it on whether the fix shipped. |
| ₹2 crore – ₹20 crore | A monthly retainer, if your buyers search and your margins hold. | This is where the four-discipline argument starts paying. It's also where most agencies coast, because you can afford not to notice. |
| ₹20 crore+ | A retainer plus an internal owner who briefs and approves. | Without the internal owner, the agency spends half its hours chasing your team for approvals. |
Reason one: the data you can't justify buying on your own
The unglamorous half of an agency's value is a licence stack you'd never sign off for a single site. A mid-tier Ahrefs or Semrush seat runs roughly ₹2–3 lakh a year at list price and prices move, and that's one tool. Add a crawler, rank tracking across enough keywords to be honest, a log-file parser, and something to watch how often your pages get cited in AI answers, and you're looking at a five-to-seven-figure annual line item before anyone does any work.
The tools alone aren't the point though — you can rent most of them monthly. What you can't rent is the aggregated read. An agency running thirty sites sees a core update land across all thirty in the same week. You see one graph and a bad feeling.
That difference matters most on the day something goes wrong. Knowing whether a 30% drop is Google reshuffling the whole category or something your developer shipped on Tuesday is the difference between a calm fortnight and a panicked rewrite of your entire site.
Reason two: pattern recognition, which mostly means avoiding expensive mistakes
Most people assume they're buying clever tactics. They're mostly buying mistake avoidance, and that's the better deal.
SEO has a small set of failure modes that cost six to twelve months each, and they all look reasonable while you're walking into them. Somebody who has watched them happen twenty times spots them in the brief, not in the post-mortem.
- A replatform with no redirect map. The traffic doesn't drop on launch day. It drops three weeks later, which is exactly when nobody connects the two events.
- Twelve pages fighting over one query. You wrote all of them in good faith. Google picks the worst one and rotates. See what keyword cannibalization is.
- A publishing cadence nobody can hold. Twenty articles in month one, silence by month four. Consistency beats volume and it isn't close.
- Links from a rate card. Cheap, fast, and a penalty with a delayed fuse.
- Optimising pages that have no commercial intent behind them. Rankings go up, pipeline doesn't move, and everyone gets quietly cynical about SEO.
Reason three: SEO is four jobs, and you can't hire a quarter of a person
Technical, content, links, analysis. Those are four different skill sets with four different day-to-day rhythms, and almost nobody is genuinely strong at more than two. The person who enjoys reading log files is rarely the person who writes a good comparison page.
That's the structural argument for an agency and it's the only one that's hard to argue with. You're buying fractional access to four specialists in a month where you need eight hours of technical work, forty hours of content and twenty of outreach — a split that would be impossible to staff internally without three people sitting idle most of the quarter. How SEO agencies actually work breaks that split down hour by hour.
The moment that argument stops holding is when your needs stabilise. If you need forty hours of content and nothing else every single month, you don't need an agency. You need a writer and an editor.
Two reasons to hire that don't survive scrutiny
Both of these get agencies hired every week in India. Both produce engagements that everyone quietly resents by month five.
- "Our competitor ranks above us and it's driving me mad." Possibly they rank for a query nobody buys from. Before you spend ₹9 lakh closing a gap, check whether the gap is worth anything: look at what the query actually is, whether the page behind it sells anything, and whether your own customers describe their problem that way. Competitive irritation is a real feeling and a terrible brief.
- "Our blog is empty and it looks bad." An empty blog costs you nothing. A blog full of posts nobody searches for costs you the retainer, plus the reviewing hours, plus the crawl budget, plus the awkwardness of eventually deleting it. If the only stated goal is volume, you're buying a content calendar, not a growth channel.
Under the threshold? Here's what actually beats a retainer
Being too small for an agency is not the same as being too small for SEO. The work is the same work; you just do the highest-yield 20% of it yourself and stop.
In rough order of return per hour:
- Claim and finish your Google Business Profile, if you serve anyone locally. Categories, hours, services, real photos, and a habit of asking for reviews. This is the single highest-return unpaid hour in local marketing.
- Fix speed and mobile rendering. Not a redesign — image compression, a decent host, and getting the layout to stop jumping around while it loads.
- Write the eight pages your buyers already ask about. Pricing. What you do and don't do. You versus the obvious alternative. The two or three use cases you're genuinely good at. These convert; "10 tips" posts don't.
- Set up Search Console and GA4 properly, then leave them alone for 90 days. When you do hire, you'll have a real baseline instead of a shrug.
- Buy a one-off audit rather than a retainer. You get the plan without the standing invoice — see what an SEO audit should actually find.
- Consider a freelancer instead. Below roughly ₹40,000 a month a good individual usually beats a cheap agency. Agency or freelancer walks through where that stops being true.
Run the payback maths before you take the first sales call
Five lines on the back of an envelope will tell you more than any agency pitch. Here's the shape of it, using ₹75,000 a month as the input. Treat the numbers as a worked example — substitute your own and the conclusion often flips.
₹75,000 × 12 = ₹9 lakh committed. Say your average deal is ₹1.5 lakh at 60% gross margin, so each closed deal is worth ₹90,000 in gross profit. Break-even is 10 extra deals in the year. If you close 20% of qualified leads, that's 50 extra qualified leads, or roughly four a month.
Now the only question that matters: looking at the queries your buyers actually type, is four extra qualified leads a month plausible inside a year? Sometimes the answer is obviously yes. Sometimes you look at the search volume for your category in India and realise the whole market is 400 searches a month, at which point SEO is a supporting act and paid social is the show.
Ask any agency to do this arithmetic with you on the first call. The ones who do it cheerfully are the ones worth hiring. The ones who redirect to traffic projections are telling you something too.