If your website gets traffic but no conversions, one of eight things is almost always wrong: mismatched traffic intent, a leaking funnel step, mobile friction, hidden technical errors, an invisible or weak call to action, missing trust signals, surprise costs, or a form that asks too much. The numbers alone can't tell you which — a 2% conversion rate looks identical whether the problem is your ad targeting or a broken checkout button. But visitor behavior can. This guide walks the exact diagnosis, step by step, with real dashboard examples at every stage.
Key takeaways
- Eight causes explain almost every "traffic but no sales" problem; the most common are wrong-intent traffic, one leaking funnel step, and mobile friction.
- Build a funnel first: it turns "conversions are low" into the single step that leaks — in our demo store, 55% of started checkouts were abandoned.
- Watch ten recordings of visitors who left at that step; the cause is usually visible within an hour.
- Split results by traffic source and device before redesigning anything — a blended conversion rate hides the real problem.
- Five modest 15% improvements compound to roughly double a conversion rate.
We'll use a demo store ("Acme Outfitters") inside Clicktics for the walkthrough: 60 sessions, 5 purchases, and a very typical story hiding in between. Every screenshot below is the actual product working on that demo data — follow along with your own site and the same workflow applies.
First, confirm what "no conversions" actually means
Before touching the site, split one number into two. Overall conversion rate = (people who wanted to buy) × (people who managed to). Low conversions mean either the wrong visitors or the wrong experience — and they're fixed in completely different places (your ads vs your website). The fastest way to tell them apart is a funnel: if visitors never even reach a product page, it's an audience problem; if they reach checkout and vanish, it's an experience problem.
Step 1: Build a funnel and find the leaking step
A funnel turns "conversions are low" into "we lose 55% of buyers at checkout" — a problem you can actually work on. Define the 4–6 pages every buyer must pass through (landing → product → cart → checkout → confirmation) and count how many sessions survive each step.
Read it like a doctor reads an X-ray. In our demo: 42% bounce before viewing a product (normal-ish), 46% of product viewers never add to cart (merchandising/price question), and then the red flag — 55% of people who started checkout didn't finish. Those are the most motivated visitors on the entire site; losing more than half of them at the last step is where the money is. The Page Performance table below the funnel adds a second clue: the checkout page's exit rate (54.5%) is the worst on the site while its average time is just over a minute — people aren't deliberating, they're hitting something and leaving.
Step 2: Watch recordings of the visitors who left
The funnel tells you where; recordings tell you why. Filter session recordings to the exact segment that leaked — in our case, sessions that reached /checkout but never hit the confirmation page — and watch ten of them end to end.
Open a recording and you see the visit replayed like a video — every scroll, click, hesitation and retry — with the visitor's journey on a timeline beside it:
Patterns emerge brutally fast. Across ten abandoned checkouts you'll typically spot one of these within the hour: a payment field that rejects valid input, a coupon box that errors, shipping costs appearing for the first time, a "create account" wall, or — on mobile — a keyboard covering the pay button. You're no longer guessing; you're watching the reason.
Step 3: Check whether anyone actually sees your call to action
If the funnel shows the leak is earlier — plenty of traffic, almost nobody clicking through to products — the question becomes attention. A click heatmap stacks every visitor's clicks on top of the real page, so you can see in one picture what people engage with and what they scroll past.
Three heatmap patterns explain most traffic-no-conversion sites: clicks clustered on the navigation while the hero CTA stays cold (visitors are hunting, not converting); heavy clicks on images or headings that aren't links ("dead clicks" — expectation mismatch); and heat that dies before the fold where your signup section lives. Each has a one-line fix: move the CTA up, make the clicked thing clickable, shorten the page.
Step 4: Segment — because your average is lying to you
A blended conversion rate hides its own cause. Split it two ways and the problem usually falls out immediately:
- By source. In the demo data you can already see it: Instagram and Facebook visitors land, view one page for ~40 seconds and exit; email and direct visitors walk four pages deep to /checkout. If paid social is most of your traffic, your "conversion problem" is an intent problem — the site may be fine. Fix the targeting or build a dedicated landing page; don't redesign the homepage for an audience that was never buying.
- By device. Desktop converting at 4% while mobile sits at 0.8% is not a conversion problem, it's a mobile problem — and with 60%+ of traffic on phones, it quietly halves your blended rate. Watch five mobile recordings of your checkout and you'll usually find the culprit in minutes.
Step 5: Hunt the silent killers — errors nobody reports
Some visitors don't choose to leave; your site breaks and ejects them. A JavaScript error on one browser version, a payment script that fails behind an ad-blocker, a form that silently rejects valid phone numbers — your team never sees these because your team's devices work. This is why error tracking belongs inside your analytics: Clicktics captures console errors tied to the exact session where they happened, so a spike in errors on /checkout comes with the recordings of every visitor who hit it. One fixed bug here routinely outperforms months of copy testing.
Step 6: Price the leak before you fix anything
Not every drop deserves engineering time, so put a dollar figure on each one before choosing. The arithmetic takes two minutes. Scale the demo funnel to a real store doing 20,000 visitors a month at an $80 average order: with the same shape, roughly 11,600 visitors view a product, 6,400 reach the cart, 3,600 start checkout, and — at that 45% completion rate — about 1,620 purchase, for $129,600 in monthly revenue.
Now price the candidate fixes. Lifting checkout completion from 45% to a merely-average 65% (the kind of jump that fixing one broken payment field or removing a forced account wall produces) takes purchases from 1,620 to roughly 2,340 — about $57,000 of additional monthly revenue from one step. Compare that with heroically improving the landing page's product click-through by two points, worth perhaps $4,000–6,000 a month, and the priority order writes itself. This is why the funnel comes first in the workflow: it doesn't just locate problems, it ranks them by payroll impact — and it's why "redesign the homepage" is so rarely the right first move.
The same math protects you from over-celebrating, too. If a fix "feels" better but the step's conversion doesn't move within a full week of traffic, the funnel will say so, and you move to the next candidate instead of shipping cosmetics for a month.
Four mistakes that waste a perfectly good diagnosis
- Redesigning instead of repairing. A full redesign resets everything at once — you'll never know which change mattered, and you risk breaking the steps that worked. Fix the one leaking step; leave the rest alone.
- Watching only converted sessions. They're satisfying, but they teach you nothing about the leak. The information lives in the sessions that failed — filter for them deliberately.
- Trusting a week of gut feel over a week of data. Behavior data accumulates fast at even modest traffic; opinions don't improve with volume. Let the before/after step-conversion numbers decide.
- Ignoring the "boring" causes. Teams love testing headlines and hate checking JavaScript errors — yet the error feed fixes more revenue per hour spent than any copy test. Do the boring sweep first.
The 8 causes, ranked by how often they're the answer
| # | Cause | How you spot it | The fix |
|---|---|---|---|
| 1 | Wrong-intent traffic | One source bounces in seconds while others convert (visitor table) | Retarget ads, match landing page to ad promise |
| 2 | One leaking funnel step | A single −50%+ drop in the funnel | Fix that step only — ignore the rest for now |
| 3 | Mobile friction | Mobile converts at a fraction of desktop | Watch mobile recordings of checkout; fix forms/keyboard/tap targets |
| 4 | Hidden technical errors | JS errors clustered on money pages | Fix the script; verify with new recordings |
| 5 | Invisible or weak CTA | Heatmap cold on the primary button | Move it above the fold, one clear action per page |
| 6 | Missing trust signals | Recordings show pricing-page hesitation, exits to search engines | Reviews, guarantees, security badges near the CTA |
| 7 | Surprise costs | Cart→checkout drop; recordings show exits when shipping appears | Show total costs early; free-shipping threshold |
| 8 | Forms that ask too much | Recordings show field-by-field abandonment | Cut fields, allow guest checkout, autofill |
What this looks like as a weekly routine
- Monday (20 min): open the funnel, note the worst step vs last week.
- Tuesday (30 min): watch 10 recordings of that step's abandoners; write down every point of friction you see twice.
- Wednesday: ship the smallest fix on the list. No redesigns — one change.
- Rest of week: let data accumulate; check the error feed for anything new on money pages.
- Next Monday: compare the step's conversion against last week. Keep what worked; repeat.
Teams that run this loop typically fix their first real leak in week one — because the tooling removes the guessing, not because the fixes are clever. Compounding does the rest: five modest 15% improvements roughly double a conversion rate, a calculation we walk through in our conversion playbook.
Frequently asked questions
Why am I getting traffic but no sales?
Most often: the traffic's intent doesn't match the page (especially paid social), a single funnel step is leaking badly, or mobile checkout friction is dragging the blended rate down. Build a funnel to locate the leak, then watch recordings of the visitors who abandoned there — the cause is usually visible within ten sessions.
How much traffic should convert?
Benchmarks vary hugely by industry — ecommerce typically converts 1.5–3% of visitors, SaaS trials 2–7% (full numbers in our conversion rate by industry reference). If you're well below your industry's floor, suspect a specific defect rather than general "optimization" needs.
What tools do I need to diagnose this?
Three capabilities: a funnel (to find the leaking step), session recordings (to see why), and heatmaps (to check attention) — plus error capture for the silent failures. Clicktics includes all four in one ~24 KB script with a 14-day free trial, which is how the walkthrough above was produced.
Is low conversion always the website's fault?
No — and that's the first thing to rule out. If one traffic source converts fine while another doesn't, the site works; the audience is wrong. That's an advertising fix, not a design fix, and it's why segmenting by source is Step 4 instead of an afterthought.
How long until I see improvement?
Recordings start revealing friction the first day. If the cause is a defect (bug, surprise cost, broken field), fixing it moves numbers within a billing cycle. If it's traffic quality, improvements track your ad changes. The diagnosis itself — knowing which — takes about an afternoon with the workflow above.
Find your leak this week
Your traffic already paid for the answers — every abandoned session is a recorded explanation of what went wrong. Clicktics gives you the funnel, the recordings, the heatmaps, the visitor journeys and the error capture from one ~24 KB script, with a 14-day free trial and no credit card. Install it today; by next week, "traffic but no conversions" becomes a short list of fixes with your name on it.
Start your 14-day free trial → or see the full feature tour.
Tomás García
Tomás García writes for the Clicktics blog about session replay, analytics engineering, and building privacy-first products that agencies love. Reach the team at [email protected].