What Is a Good Conversion Rate? The Honest Answer

Why the benchmark question is the wrong one

Published benchmarks are real. Most B2B lead generation falls somewhere between 2% and 5%; ecommerce mostly between 1% and 3%. Those numbers are accurate and almost entirely unusable, for one reason: the spread inside any industry is far wider than the gap between industries.

Two businesses selling the same service in the same city can convert at 0.8% and 6% because one runs brand search and the other runs cold social traffic. That difference has nothing to do with the industry average and everything to do with who is arriving and what they were promised on the way in.

So a benchmark answers “am I normal?” when the useful questions are “where is my traffic actually leaking?” and “what would fixing it be worth?” Neither has an industry answer.

The only benchmark that tells you anything is your own number last month, on the same page, from the same channel.

Calculating it properly

Conversions ÷ visitors × 100. The arithmetic is trivial. The definitions are where it goes wrong.

Sessions or users?

Sessions is the usual choice and what GA4 reports by default. Users produces a higher rate because one person visiting three times counts once rather than three times. Both are defensible; mixing them is not, and it is a common reason two people in the same company quote different numbers for the same page.

Same period, both numbers

Conversions from this month against visitors from last month is not a conversion rate, and it happens more often than you would think when two reports are pulled from two tools with different default ranges.

What counts as a conversion?

Decide once and write it down. Form submissions only? Phone clicks and WhatsApp clicks too? A newsletter signup? Each is defensible and they produce very different numbers. The failure is a definition that quietly changes when somebody adds a new event, so that a step-change in the graph looks like a marketing result and is an instrumentation change.

Why the site-wide number lies

A single conversion rate for a whole site is an average across pages doing completely different jobs.

1.4% describes none of these pages. Worse, it moves whenever the traffic mix moves — a successful blog post drags the site-wide rate down while adding real value, and somebody looking only at the headline number will conclude that content marketing is hurting conversion.

Calculate per page, or at minimum per channel and per landing page group. It is more work once and it stops you drawing conclusions from an artefact of your own traffic mix.

Setting a target that is a number

“Improve conversion” is not a goal. It has no size, so nobody can tell whether it was achieved or whether the effort was worth it.

Convert it into arithmetic instead. At 4,000 visitors and 60 conversions you are at 1.5%. Moving to 2% means 80 conversions — 20 more, at the same traffic. If a customer is worth ₹15,000 and one in four enquiries closes, those 20 extra enquiries are ₹75,000. Now the question “is a week of work on this page worth it?” answers itself.

This reframing does two useful things. It makes the goal falsifiable, and it makes the comparison against buying more traffic explicit — because 20 extra conversions from paid media has a price you can look up, and it is usually much higher than the cost of fixing a page.

What actually moves it

In rough order of how often the fix works, on the pages I look at:

1. Stating the offer plainly, high on the page. Astonishingly many pages never say in ordinary words what is being sold and to whom.
2. One obvious next step, above the fold and repeated. Covered in how many CTAs a page should have.
3. More doors. Form, phone and WhatsApp convert different people. A site with one door loses the rest.
4. Shorter forms. Every additional field costs completions. Ask for what you need to reply, not what you would like to know.
5. Proof. Testimonials with real names, results with real numbers. Trust failures look identical to disinterest in your analytics.
6. Speed. Not glamorous, and it disproportionately affects mobile visitors on slow connections — who are frequently the majority.
7. Traffic quality. Last, deliberately. It is the most common first explanation and it is usually not the constraint until the six above are done.

Why a calculator with a target table

The division is something you could do on a phone. What the calculator adds is the part people skip: the target table. Enter your visitors and conversions and it shows each higher rate alongside exactly how many additional conversions it would require at your current traffic — and, if you enter a value per conversion, what that gap is worth.

That table is what turns “we should improve conversion” into “we need twenty more enquiries a month, worth ₹75,000, and here is what that costs to attempt”. It is the single most useful output, and almost no calculator provides it.

It also shows the sample size next to the rate, because nine conversions from 400 visitors is 2.25% and means considerably less than the number suggests. Knowing when a movement is real and when it is noise is a large part of not wasting months, and it has its own post.

Free, no signup, and it runs entirely in your browser — nothing you type is uploaded or stored. Once you know the number, the ROI calculator turns it into cost per acquisition and return, and the CTA checker looks at whether the page is even asking.

FAQ