Q3
If your site converts at 1% and you double your traffic, you've doubled the number of people leaving without buying. Conversion funnel optimization is the work of fixing the leaks first, so every visitor you pay for afterwards is worth more. It's almost always cheaper than buying more traffic, and the gains compound across every channel you run, because a better conversion rate improves your paid ads, your organic search and your email at the same time.
Here's the arithmetic that should decide your budget. A site getting 10,000 visits a month at 1% conversion produces 100 sales. Getting to 2% produces the same 100 extra sales as buying another 10,000 visits, except the rate improvement is permanent and the traffic is not. You pay for traffic every month. You pay to fix a broken checkout once.
Why more traffic feels like the answer
Because it's easier to buy. Traffic has a price list, a dashboard, and someone willing to sell it to you this afternoon. Funnel work means auditing your own pages and admitting that something you built is losing money.
It's also easier to measure badly. Sessions go up, the chart looks healthy, and nobody notices that revenue moved less than traffic did. That gap between traffic growth and revenue growth is the clearest sign your funnel is leaking, and it's the first thing we look at in any account.
There's a third reason. Agencies are usually paid to deliver traffic. Very few are paid to reduce the amount of it you need.
What the research says about where money leaks
Ecommerce has the best public data on this, and the numbers are blunt. Baymard Institute's aggregate of 50 studies puts the average cart abandonment rate at 70.22%, a figure that has barely moved in almost twenty years.
The useful part isn't the headline. It's the split. Baymard's survey found 42% of US online shoppers abandon because they were just browsing and not ready to buy. That group was never yours to win. Strip them out and the remaining reasons look like this:
| Reason for abandoning | Share | Fixable? |
|---|---|---|
| Extra costs too high (shipping, tax, fees) | 40% | Yes |
| Delivery was too slow | 20% | Partly |
| Didn't trust the site with card details | 19% | Yes |
| Site wanted them to create an account | 18% | Yes |
| Checkout too long or complicated | 17% | Yes |
| Website had errors or crashed | 17% | Yes |
| Returns policy wasn't satisfactory | 13% | Yes |
| Couldn't see total cost up front | 12% | Yes |
| Card was declined | 10% | Partly |
| Not enough payment methods | 9% | Yes |
Respondents could pick more than one, so the column doesn't total 100%. Look at how many of those are design and policy decisions rather than market forces. Surprise costs, forced account creation, hidden totals, a checkout with too many steps. Google's own page experience guidance makes a related point: what frustrates users and what search rewards largely overlap. None of those get better with more traffic. All of them get worse.
Baymard also quantifies the opportunity. Across large US and EU ecommerce sites, they find an average 35.26% conversion rate increase is achievable through checkout design alone, which on combined sales of $738 billion works out to roughly $260 billion in recoverable orders. Their benchmark of 60 leading sites found the average one still has 39 separate areas needing checkout improvement. These are companies with dedicated optimisation teams.
One specific finding is worth sitting with. Baymard's testing shows a good checkout can run on 12 to 14 form elements, yet the average US checkout displays 23.48 by default. Most sites could cut 20% to 60% of their form fields without losing a single piece of information they actually need.
This applies to service businesses too
If you sell services, you don't have a cart, so you assume none of this is about you. It is. Your funnel is just less instrumented.
A typical service funnel runs: ad or search result, landing page, service page, contact form, form submission, reply, call booked, proposal sent, client. Most businesses measure the first step and the last step and nothing in between. Which means when enquiries drop, you can't tell whether fewer people visited, more people bounced, the form broke, or your follow-up got slow.
The service-side equivalents of Baymard's ecommerce leaks:
- Extra costs too high becomes no pricing guidance anywhere on the site, so unqualified enquiries flood in and good prospects assume they can't afford you
- Forced account creation becomes a form asking for company size, budget and phone number before you've earned any of it
- Didn't trust the site becomes no named team, no real work shown, no address
- Checkout too long becomes a four-step booking flow to get a 15 minute call
- Site had errors becomes a contact form that silently fails on mobile, which is more common than anyone wants to believe
Test your own form today. Fill it in from your phone and confirm the email actually arrives. We find broken forms more often than we find bad copy.
How to find your leaks in an afternoon
A single conversion rate tells you nothing about where people leave. You need step-by-step drop-off. GA4's funnel exploration does this and most businesses have it sitting unused.
A few mechanics that catch people out. You can define up to 10 steps. An open funnel lets users enter at any step, a closed funnel only counts those who start at step one, and if someone skips a step they drop out and aren't counted in any later step. Turn on "Show elapsed time" to see average time between steps, which surfaces hesitation you can't otherwise see. And use the breakdown dimension for device category, because mobile and desktop funnels usually fail in different places.
Build these steps for a service site: landing page view, service page view, contact page view, form start, form submit. For ecommerce: product view, add to cart, begin checkout, payment info, purchase.
Then read it like this:
- Find the single biggest percentage drop between two steps. Ignore the step with the fewest users. You want the step with the steepest fall.
- Break that step down by device. If mobile is much worse, you have a layout or speed problem, not a persuasion problem.
- Go and use that step yourself, on a phone, on mobile data, as a stranger would.
- Write down what's confusing or slow. Your list will be short and mostly obvious.
- Fix the cheapest item first, not the most interesting one.
That's the whole method. Conversion funnel optimization gets sold as a specialist discipline with expensive tooling, and the tooling helps at scale, but the first 80% is watching where people fall over and removing the obstacle.
What to fix first, ranked by effort against payoff
Work down this list. Most businesses find something in the first three rows.
| Fix | Effort | Typical impact |
|---|---|---|
| Show total costs, including shipping and tax, before checkout | Low | High |
| Remove required form fields you don't use | Low | High |
| Offer guest checkout, or a no-account enquiry path | Low | High |
| Fix mobile layout and speed on the worst-performing step | Medium | High |
| Add trust signals where the money question happens | Low | Medium |
| Publish pricing guidance or a starting range | Low | Medium |
| Shorten multi-step flows into fewer screens | Medium | Medium |
| Add more payment methods | Medium | Medium |
| Rewrite headlines and calls to action | Low | Low to medium |
| Redesign the whole site | High | Unpredictable |
Note where a full redesign sits. It's the thing businesses reach for most and the thing with the least reliable return, because it changes everything at once and you learn nothing about which change helped.
When more traffic actually is the answer
Being honest about the limits of this argument: sometimes you really do need more traffic.
If you're converting at a healthy rate for your industry and simply don't have enough visitors to hit your revenue target, optimisation has less room to work. Squeezing 2.5% up to 2.8% on 400 visits a month is not a business plan. You need volume, and you should go buy it.
The same applies to a new site with no traffic at all. There's no funnel to optimise yet, and no data to optimise it with. Get visitors first, then read what they do.
The test we'd use: if you have at least 1,000 relevant visitors a month and a conversion rate below your industry norm, fix the funnel first. Above the norm, or below that traffic level, buy traffic and instrument it properly while you do.
Where to start this week
Pick your single highest-value conversion, the one that pays the bills. Build a funnel for it in GA4 with five steps or fewer. Find the biggest drop. Then complete that step yourself on your phone and write down everything that annoys you.
That exercise takes about two hours and will give you a shortlist worth more than a month of extra ad spend. If you want help running it, our landing pages and CRO service starts with this audit, and analytics and reporting covers the measurement setup that makes it repeatable, and the CRO guides go deeper on individual steps. A free site audit with Buzzing Bee will also flag the technical problems, slow pages and broken elements, that quietly widen every leak you have. Before you increase any budget, run the numbers through our SEO ROI calculator to see what a rate improvement is worth against the cost of more visitors.
FAQs
What is conversion funnel optimization?
It's the practice of measuring each step a visitor takes toward becoming a customer, finding where most of them drop out, and removing whatever causes it. The goal is a higher percentage of existing visitors converting, rather than more visitors. Gains apply to every traffic source at once, so the effect compounds.
How do I know if my funnel is leaking?
Compare your traffic growth against your revenue growth over the same period. If sessions rose 40% and revenue rose 10%, you're leaking. Then build a funnel report in GA4 and look for any single step losing a disproportionate share of users compared with the steps around it.
What is a good conversion rate?
It varies so widely by industry, price point and traffic source that benchmarks are close to useless for decisions. A better target is your own trailing three-month average. Beating your own baseline is measurable and real. Chasing an industry average sourced from someone else's business usually is not.
How much can I realistically improve conversions?
For a site that has never been optimised, gains in the tens of percent are common, because the early fixes are things like surprise shipping costs and unnecessary form fields. For a site already optimised repeatedly, expect single-digit gains that take more work. Baymard's research suggests even large, well-resourced sites still have substantial room.
Should I fix the funnel or run ads first?
If you have reasonable traffic and poor conversion, fix the funnel, because ads into a leaky funnel waste a fixed percentage of every pound you spend. If you have almost no traffic, run ads first and instrument them carefully. The ads then generate the data you need to optimise.
Do I need expensive testing tools for this?
Not to start. Analytics you already have, plus using your own site on a phone, will find your biggest problems. A/B testing tools become worthwhile once the obvious issues are fixed and you have enough traffic for tests to reach significance, which most small sites do not.