Ads & CRO

I run performance advertising and conversion-rate optimisation as one discipline: traffic you pay for, landing on pages engineered to convert it, measured end to end so every rupee reports for duty.

Why these two are one job

The standard arrangement fails predictably. An agency optimises clicks. Nobody owns what happens after the click. The landing page is the homepage, with its eleven exits and its general-purpose message. Cost per click rises because everyone in the category is bidding, conversion stays flat because nothing downstream changed, and the quarterly review becomes a discussion about creative.

Ads and CRO are the same problem viewed from either end. Splitting them across vendors is how the hole in the bucket stays unowned — the ads vendor points at the page, the web vendor points at the traffic quality, and the spend continues. Paisa lagana easy hai; paisa vasool is engineering.

What the engagement covers

• Account architecture. Campaigns mapped to intent tiers rather than to whoever built the account last. Search intent, competitor intent and remarketing behave differently and deserve different structures, budgets and expectations.
• Purpose-built landing pages. One page, one offer, one action. Built to load fast on a phone on mobile data, because that is where the traffic actually arrives.
• Tracking that survives audit. Conversions defined as revenue events rather than button clicks, server-side where the browser cannot be trusted, and offline-conversion import for B2B so the platform learns from closed deals rather than from enquiries.
• Systematic CRO. Hypothesis, sample size computed before starting, test, verdict recorded — keep, kill or learn. No carousel redesigns justified by best practice.
• Weekly numbers you can read in a minute. Spent, made, verdict.

The discipline that makes the difference

Sample size before the test, not after. A variant that looks thirty per cent better after two days is usually noise, and calling it produces a documented win that quietly fails to appear in revenue. Do that ten times and you have a folder of successes and a flat conversion rate — a genuinely confusing position that leads teams to conclude CRO does not work, when what did not work was the method. The full argument is here.

Revenue-defined conversions. A form fill is not a conversion, it is a hope. Where the sales cycle allows, closed revenue is imported back so the bidding algorithms optimise toward customers rather than toward form-fillers. This single change often reshapes which campaigns look profitable.

One hypothesis per test. Changing five things and measuring one number tells you the bundle worked, not which part did — and next time you will carry all five forward, including whichever one was harmful.

What I will tell you to stop doing

Bidding on your own brand name when nobody else is — you are paying for clicks you would have received free. Running broad match without negative-keyword discipline. Sending every campaign to one page. Judging a channel on last-click attribution when your buying cycle is three months long. And optimising toward whichever metric sits nearest to hand rather than the one you actually sell.

None of that requires a consultant to notice. It requires someone whose fee does not rise with your spend to say it out loud.

The first-call audit

Bring last month’s advertising report. I will find where the bucket leaks — usually within the first fifteen minutes, usually in tracking rather than in creative — and tell you whether the fix is worth my involvement or whether you should simply do three specific things yourself.

That is free, and a meaningful share of those calls end with the second answer.

Where the money usually leaks

Five places, in the order I find them.

Tracking that lies. Conversions counted twice, or counted on a thank-you page people reach by other routes, or lost entirely on mobile Safari. Until the measurement is trustworthy, every optimisation decision above it is guesswork with a spreadsheet. This is the first thing I check and it is broken more often than not.

Attribution mismatched to the sales cycle. Judging a three-month B2B cycle on last-click gives credit to the final branded search and none to the channel that created the demand. Businesses routinely cut the channel that was actually working because the model could not see it.

One landing page for many intents. Someone searching for a price and someone searching for a comparison want different pages. Sending both to the same generic page halves both.

Match types without discipline. Broad match with a thin negative list quietly funds an education for the algorithm at your expense.

Bidding on brand when nobody contests it. Paying for clicks that were already yours, defended with a plausible-sounding argument about competitors who are not actually bidding.

The reporting philosophy

One page, weekly, three numbers: what went out, what came back, and what I intend to do about it. Underneath sits everything — campaign structure, test log, page performance — available whenever you want to interrogate it.

The reason for the one page is not simplicity for its own sake. It is that a report which takes twenty minutes to read does not get read, and a report that does not get read cannot be challenged. Being challenged is the point.

The same logic governs the test log: every experiment recorded with its hypothesis, its sample size, its result and its verdict, kept in a document you own. When the engagement ends you keep the accumulated knowledge of what does and does not work on your audience, which is worth considerably more than the campaigns themselves. Agencies rarely hand that over, because the accumulated learning is what makes leaving them expensive.

Who this is not for

Businesses below the spend threshold where tests can resolve, brands wanting awareness campaigns measured on impressions, and anyone who needs a media agency with a creative studio attached. In each case I will name the mismatch on the first call rather than taking a month’s fee to discover it together.

Related reading

/blog/conversion-optimisation — the method in full · /services/marketing — where paid fits in the wider system · /services/website-design — the pages traffic lands on · /blog/local-seo — the organic counterpart for local demand

Bring the report

/contact — last month’s numbers, thirty minutes, seedhi baat about where the money is going.