Most conversion rate benchmarks are worse than useless, because they get quoted without the one detail that makes them mean anything: what was being counted, and for who. A 2% conversion rate is a disaster for a booked-call page and a strong result for a cold-traffic ecommerce store. Same number, opposite verdict.
This is the framing we use when a client asks whether their conversion rate is good. The answer is almost always the same. The benchmark matters less than the trend, and the trend only matters once you’ve checked that the number is real.
Everything below is a conversion rate, not a click-through rate. Click-through rate is who clicked the ad. Conversion rate is who arrived and then did the thing you wanted: bought something, or filled in the form. Conversions over visits, nothing more complicated than that.
Why most benchmarks fall apart on contact
Published benchmarks average across wildly different businesses. An “ecommerce average” folds a R400 impulse candle into a R40,000 sofa. Those two products can’t share a conversion rate, and no amount of averaging makes the blend useful to either one.
Three things break a benchmark before you can apply it.
- Traffic mix. Branded and returning traffic converts far better than cold traffic. Contentsquare’s 2026 benchmark found 2.9% of returning visitors’ sessions converting against 1.7% of new visitors’, so a site with a large repeat audience will beat a site running mostly cold traffic even when the second site is better built.
- Price and deliberation. Higher price means longer consideration, which means lower session conversion and a longer path. That isn’t a CRO problem, it’s how buying works.
- What counts as a conversion. Newsletter signup, add to cart, form submit, booked call, closed deal. Two teams counting two different events will report two different conversion rates for the same month, and both will be right.
Treat any single published figure as a starting hypothesis. The useful version of a benchmark is a range attached to a page type and a traffic source.
What the published data actually says
Four sources, each with the dataset attached so you can judge whether any of it applies to you.
Landing pages: goal completions per visit. Unbounce’s Conversion Benchmark Report, built on 41,000 landing pages, 57 million conversions and 464 million pageviews, puts the median landing page at 6.6%. The conversion here is whatever that page was built to do, usually a form submission or a signup. By industry the medians run: SaaS 3.8%, ecommerce 4.2%, travel and hospitality 4.8%, professional services 6.1%, legal 6.3%, financial services 8.3%, education 8.4%, entertainment 12.3%. These are dedicated landing pages, not whole websites.
Ecommerce, sitewide: completed purchases per session. IRP Commerce tracks the ecommerce market monthly and states its formula outright: transactions divided by sessions. The all-market average was 2.23% in August 2026, up from 1.85% a year earlier. The spread by sector is wide: Arts and Crafts 5.81%, Sports and Recreation 2.11%, Fashion 1.86%, Food and Drink 1.58%.
Shopify stores: purchases per session. Littledata’s benchmark of 2,800 Shopify sites puts the average at 1.4%, the top 20% above 3.2%, and the top 10% above 4.7%. Same action as the IRP figure, narrower population, and roughly a third lower.
By device and channel: converting sessions. Contentsquare’s 2026 report, built on 99 billion sessions, found desktop converting 74% more often than mobile. Paid search sessions came in at 2.8%. A conversion here means a purchase in retail and an application in finance, so the blend is doing some work.
Only IRP publishes its formula. The other three report a conversion rate without saying exactly what they divided by what, which is the gap this whole article is about. Quote them as indicative, never as a target to hit.
Look at the gap between the first two figures. A 6.6% median landing page and a 2.23% average ecommerce store differ by roughly threefold, and neither one is broken. One counts a form fill on a single purpose-built page. The other counts money changing hands across a whole store. So when someone tells you your 2.1% is below average, the first question is which of those datasets they were reading off.
Two numbers the published reports miss
Public benchmarks cluster around the metrics that are easy to collect. Two of the numbers we care most about barely appear, so the ranges below are our own operating figures, not industry data.
- Booked-call page, qualified traffic (visitors who book a call): 15% to 30%. High, because the visitor already decided to talk to you somewhere upstream. This page only has to not get in the way.
- Lead to qualified opportunity, B2B services (leads that become a real opportunity): 20% to 40%. This is where most agencies stop reporting, and where the money actually is.
That second one deserves more attention than it gets. A brand doubling landing page conversion from 4% to 8% while pushing unqualified leads into the pipeline has made the sales team slower, not the business bigger. We’ve watched accounts celebrate a cost per lead drop that came entirely from people who were never going to buy. If your lead-to-opportunity rate falls while your page rate climbs, the page got easier rather than better, and the fix is usually to add friction back: a budget question, a timeline field, a clearer statement of who the offer isn’t for. Conversion rate is a means to revenue, so any benchmark you track at the top needs a matching one further down.
The denominator problem
Before you optimise anything, check the number is real. This is where most conversion audits end early, because the measurement turns out to be wrong.
Four failure modes we find over and over:
Bot and internal traffic inflate the denominator. Uptime monitors, scrapers, and your own team hitting the site from the office all count as sessions. On a low-volume site that can move the rate by a third.
Duplicate conversion events inflate everything. A form firing both a native event and a GTM tag counts every lead twice. Conversion rate looks great, cost per lead looks great, and neither is true.
Cross-domain tracking breaks attribution. A booking tool on a subdomain without linker configuration turns one visitor into two, halving your apparent conversion rate overnight.
The sitewide rate hides the page that’s failing. An overall 2.4% can be a 6% landing page and a 0.3% product page sitting in the same average. The average tells you nothing about which one to fix.
If you haven’t checked these four, your conversion rate is an opinion. Fixing the measurement is the first thing our tracking and analytics work does, and it’s worth running before any CRO work starts.
Your own baseline beats any published number
Published benchmarks answer a question you don’t have. You don’t need to know what the average South African DTC brand converts at. You need to know whether this month beats last month, at the same traffic mix, for the same offer.
Build the baseline like this:
- Segment by traffic source. Paid social, paid search, organic, direct, email. Never make decisions off the blended number.
- Segment by device. Littledata puts Shopify mobile at 1.2% against 1.9% on desktop, so mobile landing somewhere around two thirds of desktop is normal. If your mobile rate is at 20% of desktop, you’ve found a bug, not a benchmark gap.
- Segment by landing page. The page that receives the click is the unit of optimisation.
- Hold it for four weeks minimum. Weekly rates on modest traffic swing enough to fool anyone. A page getting 500 visits a month needs a big change before the difference means anything.
- Record the offer alongside the number. Change the offer and you’ve changed the experiment, so the old number stops being comparable.
Twelve weeks of segmented data is worth more than any published report. You’ll know what normal looks like for your business, which is the only way to tell a real improvement from a good week.
What to do when you’re below range
Assume the measurement is clean and the rate is still low. The order of operations matters here, because most teams start at the bottom of this list.
Check message match first. The ad promised something specific. Does the page headline repeat that promise in the same words? Mismatch is the most common reason a landing page converts at 1% when it should do 8%, and it costs nothing to fix. It sits at the top of the landing page checklist we run before spending a cent on ads for the same reason.
Then check the size of the ask. Seven form fields where three would do. A “request a quote” button where the visitor wants to see pricing. A demo booking offered to someone who came to read a comparison. The ask has to match how warm the traffic is, and most pages ask for too much too early.
Then check load speed and layout on a phone, because most of your traffic is mobile. If the hero pushes the button below three scrolls, or the page takes four seconds to paint, you lose people before they read a word of your copy.
Run tests last. Button colour and headline variants are the smallest lever you have, and testing them before the first three are right just measures noise at high cost.
A rising cost per lead often traces back to a conversion problem rather than an auction problem, which is why we treat paid media and conversion as one system rather than two separate projects.
Want to know where your conversion rate actually sits?
We run a conversion audit that starts with measurement integrity, segments your real rates by source, device and page, then tells you which single page is capping your revenue. See how our Websites and CRO system works, or book a call and we’ll go through your numbers together.