A conversion optimisation agency in Johannesburg that starts with evidence

Most websites are changed on opinion, and the loudest opinion in the room usually wins the argument. That habit is expensive, because a change nobody measured is indistinguishable from no change at all. Baymard Institute reports that the average large ecommerce site could gain roughly 35% in conversion rate through better checkout design alone. Finding gains of that kind, and proving them, is the entire job of a conversion optimisation agency.

Firewater is a conversion optimisation agency in Sandton, Johannesburg, working on the revenue hiding inside traffic you already pay for. We run research, experimentation and measurement programmes on sites we built and on sites built by somebody else entirely. The work connects to website design and development, to online store builds, and to the organic search programme feeding the top of the funnel. Nothing here gets changed on instinct alone.

The method is unglamorous, and that is rather the point of it. We find where users struggle, size the problem in money, change one thing at a time, and measure what happened afterwards. Where traffic is too low for a valid test, we say so plainly and use qualitative evidence instead of pretending the maths works. Across 26 years of client work, discipline has consistently beaten a longer list of clever ideas.

Research before testing

Testing without research produces expensive coin flips, and a year of them produces nothing you can bank. We begin with analytics, session recordings, heatmaps, on-site surveys and the questions your sales team answers every single day. Recordings show hesitation that a funnel report only hints at, such as a form field being retyped three times before someone gives up. Nielsen Norman Group found that five users uncover about 85% of usability problems, which keeps moderated testing cheap and quick.

Firewater turns that evidence into a ranked list of hypotheses, each with an expected impact and an honest estimate of the effort involved. Every entry states the observed problem, the proposed change and the single metric expected to move. The list is reviewed with your team, because commercial context and stock reality often reorder it. Research also catches the fixes that need no test at all, such as a payment step failing on one popular mobile browser.

Testing that holds up

An A/B test is a statistical exercise, and a great many reported wins fall apart under scrutiny. Tests need a sample size calculated before launch, a run time covering full business cycles, and a significance threshold agreed in advance. Stopping a test the moment the graph looks positive is the most common way agencies manufacture results for a report. Nielsen Norman Group makes a similar case for reading quantitative results with care and pairing them with qualitative work.

Here is the part usually left out of the pitch: most tests do not win. Firewater expects a meaningful share of experiments to finish flat or negative, and a programme without those is a programme being reported selectively. A flat result still kills a bad idea cheaply, and a negative result protects revenue you would have lost by launching it. Every test is logged in your reporting dashboard with the hypothesis, the outcome and the decision taken, because a conversion optimisation agency should be auditable.

Forms and checkout

Forms and checkouts are where most of the money is won, because the people reaching them have already decided to act. We cut forms back to the fields you genuinely use, validate input as it is typed, and put error messages beside the field that caused them. Checkout work covers guest ordering, delivery costs shown early, trusted payment options and an order summary the buyer can check without losing progress. On mobile, keyboard type, field order and button placement shift completion rates more than a full redesign usually does.

Firewater applies the same rigour to lead forms, since a shorter form producing unqualified enquiries helps nobody in the sales team. We measure lead quality downstream, through to closed business wherever your CRM makes that possible. The same work lowers the cost per acquisition on your Google Ads campaigns without adding a cent to the media budget. For a business already buying traffic, that is usually the fastest return available.

What improves, and how you know

You get more revenue and better qualified leads from traffic you are already paying for every month. You also get a documented record of what was tried, what worked, and what was ruled out with evidence. That record compounds, because each cycle of work starts from better information than the one before it. Media spend becomes far easier to defend internally when the conversion rate is moving in the right direction.

Firewater has worked with consumer brands including Cuddlers and Remington, in categories where a small change to a product page moves real volume.

Fill in this form to get in touch

Fill in this form
to get in touch

Explore Firewater