Is Conversion Rate a KPI You Should Track? Yes, With Three Conditions

Straight answer: yes, track conversion rate, but never on its own and never before you have enough traffic to make the number mean anything. On its own it can go up while your revenue goes down, and I have watched that exact thing happen to a client's pipeline in real time. Pair it with revenue per visitor and lead quality, and it becomes one of the most useful numbers in your business.

The short answer people want, and why it's not enough

Most business owners ask me "should I track conversion rate" because a course, an agency, or a LinkedIn post told them it's the metric that matters most. It's a fine metric. It tells you what percentage of visitors did the thing you wanted them to do, whether that's buying, booking a call, or filling in a form. But a percentage with no context is just a number that makes you feel something, and feelings are not a strategy.

Here's the bit that rarely gets said out loud: conversion rate is one of the easiest metrics in marketing to fake improvements on. Cut a form from seven fields to two and your rate will climb. Target cheaper, broader traffic instead of qualified traffic and your rate might climb too, or it might collapse, depending on the offer. Drop your price by 20 percent and watch the rate rise while your margin quietly bleeds out. None of that means your marketing got better. It means the number moved, and a number moving is not the same as the business improving.

What happened when a client "improved" their conversion rate

A few years back I worked with a B2B software client running PPC landing pages for a demo request form. The form had seven fields: name, work email, company size, job title, phone number, budget range, and a free-text "what are you looking to solve" box. Conversion rate sat at 1.9 percent. Everyone on the team, me included, agreed that felt low, so we tested cutting it down to two fields: name and email.

Conversion rate jumped to 4.1 percent within three weeks. On paper, a huge win. In the boardroom, it looked brilliant. Then the sales team started complaining. The leads coming through were unqualified, half of them personal Gmail addresses, several from students and job seekers who'd clicked through curious about the product. The sales-qualified lead rate, which had been sitting around 40 percent, dropped to 12 percent. Bookings for the following quarter fell 18 percent even though the top-of-funnel numbers looked healthier than they'd ever looked.

We put the extra fields back, added one more (a dropdown asking what stage of buying they were at), and the conversion rate settled back down to around 2.3 percent. Sales-qualified rate climbed back to 38 percent. Bookings recovered within six weeks. The lesson wasn't "never simplify a form." It was that conversion rate without a downstream quality check will lie to you cheerfully and often.

Why sample size wrecks small businesses more than anyone admits

If your site gets 40 visitors a day and you get one sale, your conversion rate that day is 2.5 percent. Get two sales the next day and it's 5 percent. You have not doubled your performance. You have two extra sales, and with numbers that small, the swing is noise, not signal.

A rough rule I give clients: don't trust a conversion rate as a trend until you have at least 100 conversions in the period you're measuring. Below that, look at the raw count of sales or leads instead, because the percentage will bounce around in a way that tempts you into "fixing" things that were never broken, or panicking over a dip that's pure statistical noise. I've seen owners rip out a landing page that had a bad week purely because the conversion rate dropped from 3 percent to 1.8 percent on a Tuesday with 60 visitors. That's four sales instead of two. It's not a crisis. It's a small sample doing what small samples do.

How to track it, step by step

This is the version I set up for clients, not the theory version:

• Step 1: Define the conversion event precisely. Is it a purchase, a form submit, a call booked, a demo request? One event, clearly named, no ambiguity. If you want the full logic behind the calculation itself, my breakdown of the conversion rate formula covers what the number measures and where it breaks down.
• Step 2: Set up the event in your analytics. This usually means Google Analytics 4 events or a similar setup, not just relying on a "thank you page views" count, which double-counts refreshes and back-button visitors. I cover how to set these up cleanly in my piece on Google Analytics events and why they matter.
• Step 3: Segment by traffic source before you look at the overall number. A blended conversion rate hides the fact that your email list converts at 6 percent and your cold social traffic converts at 0.4 percent. Blend those and you get a mediocre-looking 2.1 percent that tells you nothing useful about where to spend your next pound.
• Step 4: Wait for a real sample. As above, aim for at least 100 conversions before you draw conclusions from a rate rather than a raw count.
• Step 5: Pair it with a quality or revenue metric every single time you report it. Revenue per visitor, average order value, sales-qualified lead rate, whatever fits your business model. Never present conversion rate alone in a report again.
• Step 6: Review monthly, not daily. Daily conversion rate for most small and mid-sized businesses is just noise dressed up as insight. Monthly gives you enough volume to see a real trend.

The number to put next to it

If I could only give a client one paired metric, it would be revenue per visitor (or revenue per session). It's calculated by dividing total revenue in a period by total visitors in that same period. It automatically punishes the kind of "improvement" that comes from cheap tricks, because if your conversion rate goes up but your average order value or lead value collapses, revenue per visitor stays flat or drops, and you see the truth immediately instead of celebrating a vanity number.

For lead generation businesses, swap revenue per visitor for cost per sales-qualified lead, or even better, cost per closed deal. A form that converts at 5 percent but produces leads your sales team can't close is worse for your business than a form that converts at 2 percent and produces leads that close at twice the rate. Conversion rate, in isolation, cannot tell you which of those two forms you're running.

Where AI has changed how I look at this metric

The tools have moved on since I started tracking these numbers by hand in spreadsheets. AI-driven analytics can now flag which segments are dragging your blended conversion rate down without you manually slicing the data, and predictive models can estimate lead quality before a human even looks at the form submission. I've written before about how AI really changes marketing strategy and where it doesn't touch it at all, and conversion tracking is a good example of both: AI makes the segmentation faster, but it still can't tell you whether a rising number is good news or a warning sign. That judgement is still yours.

If you're running app-based products alongside a website, the tracking setup gets messier fast. I've had clients assume their app conversion data was flowing cleanly into the same reports as their website data, only to find the two were sitting in separate silos for months. I go through exactly how to decide on this in my piece on whether to enable Google Analytics inside Firebase, and it's worth checking before you trust a blended conversion figure across both platforms.

When conversion rate doesn't deserve a place on your dashboard

There are situations where I tell clients to drop it, or at least demote it. If you're a brand-new site with under 500 monthly visitors, don't build a dashboard around conversion rate yet, look at raw conversions and qualitative feedback instead. If you're running a multi-step funnel where the "conversion" happens weeks after the first visit (long B2B sales cycles, high-ticket coaching, anything with a considered purchase), a single site-wide conversion rate flattens a process that has five or six meaningful stages, and you'd get more value tracking conversion rate at each stage separately.

And if your business model depends on lifetime value rather than one-off transactions, subscriptions, memberships, retainer clients, conversion rate at the point of sale tells you almost nothing about whether the business is healthy. A gym that converts 8 percent of website visitors into sign-ups but loses 40 percent of members within three months has a retention problem dressed up as a conversion success story.

My verdict after years of reporting this number to clients

Track it. Report it monthly. Never present it without a revenue or quality metric sitting right next to it. Don't trust it below roughly 100 conversions in the period. Segment it by source before you draw any conclusion from the blended figure. And accept, uncomfortably, that a rising conversion rate is sometimes the first sign something downstream is about to go wrong, not proof that everything's going right. That's the part of this metric nobody puts on the front of the dashboard, and it's the part that's saved my clients from making expensive decisions based on a number that looked great and meant very little.

If your goals tend to drift after the planning meeting, my roundup of the best OKR software compares the tools that keep them on track. If you want to see where visitors click and where they give up, my roundup of the best heatmap tools compares the options.

Frequently asked questions

Is conversion rate more important than traffic?

Neither matters in isolation. High traffic with a low conversion rate can still beat low traffic with a high conversion rate if the total number of sales and the revenue per visitor are stronger, so always look at both together with the actual revenue figure, not just the two percentages side by side.

What's a good conversion rate for a small business website?

It depends heavily on industry and traffic source, but for ecommerce, 2 to 3 percent is typical across the board, while paid search landing pages for lead generation often sit between 3 and 6 percent. Treat any benchmark as a rough guide, not a target, because your own historical data matters more than an industry average.

How often should I check my conversion rate?

Monthly for decision-making, weekly at most if you're mid-way through an active test. Checking it daily on a small or medium site almost always leads to reacting to statistical noise rather than a genuine trend, which wastes time and can push you into changes that weren't needed.

Can conversion rate go up while revenue goes down?

Yes, and it happens more often than most reports admit. It usually happens when you attract cheaper, less qualified traffic, cut friction from a form in a way that lowers lead quality, or discount a product so heavily that more people buy at a much lower margin. This is exactly why conversion rate should never be reported without a revenue or lead-quality figure alongside it.