Techbanq

Web · conversion

Stop shipping opinions about the funnel and start testing them

Every organisation has theories about why visitors leave without buying or enquiring. Conversion optimisation is the discipline of turning those theories into experiments a live funnel can prove or kill — so the site changes because the evidence said to, and the gains compound instead of washing out.

Serving Saudi Arabia from our Riyadh office and our engineering hubs in Lahore and Karachi

Shoppers moving through a modern retail concourse

Enterprise delivery credentials

  • ISO

    ISO/IEC 27001

    ISO/IEC 27001 certified

Certifications & registrations

The discipline

Why the loudest voice in the room keeps redesigning your funnel

We do not have an office in Saudi Arabia

We serve organisations there from our Riyadh office and our engineering hubs in Lahore and Karachi, travelling for the parts of an engagement that need to happen in person.

Without evidence, funnel decisions default to hierarchy: the most senior opinion ships. The homepage is rearranged because a director dislikes it, the form is shortened because a conference talk said to, a competitor’s checkout is copied because theirs presumably works. Some of these changes help, some hurt, and — this is the expensive part — nobody can tell which, because nothing was measured before or compared after. The funnel drifts, and every quarter it is redesigned by whoever argued best.

The alternative is a working ritual, not a tool purchase. Instrument the funnel until each step’s drop-off is a number someone owns. Draw hypotheses from your own visitors’ behaviour — where they hesitate, what they retype, which reassurance they look for and do not find — not from another market’s case studies. Test the change against a control, read the result honestly, and keep only what won. It is slower than shipping the redesign, and it is the only version of this work whose gains are still there a year later.

A note on scope: hotel booking funnels have their own dedicated pages on this site, because a booking engine has its own economics. This page is about every other funnel — an e-commerce checkout, a lead form, a quote request, an account opening — wherever a measurable step stands between a visitor and the thing the business wants them to do.

The loop

The cycle that turns a hunch into a kept or discarded change

Each pass through this loop either improves the funnel or improves what you know about it. The branch exists because not every funnel has the traffic for a classic split test, and pretending otherwise produces confident noise.

A loop of five stages — locate the leak, explain it with evidence, state a falsifiable hypothesis, run the controlled test, keep or kill and log — with a branch to qualitative methods when traffic cannot power a split test.

  1. 01

    Locate the leak

    The instrumented funnel names the step where the drop is worst relative to its intent — the place a fix is worth the most.

  2. 02

    Explain it with evidence

    Session recordings, form analytics and user input turn the number into a story: what visitors were trying to do when they gave up.

  3. 03

    State a falsifiable hypothesis

    A specific change, a specific metric, a predicted direction — written down before the test so the result cannot be reinterpreted after.

  4. 04

    Run the controlled test

    Variant against control, run to the sample the design demands — not stopped the first morning the graph looks pleasing.

  5. 05

    Keep, kill, and log either way

    Winners ship to everyone; losers are removed; both are recorded, because a documented dead end is the cheapest insurance against re-arguing it next year.

Always available

When traffic cannot power a split test

Lower-traffic funnels switch method, not standards: moderated usability sessions, before-and-after with guard metrics, and fixing outright defects that need no test. What we do not do is run an underpowered experiment and report its noise as a finding.

Worked example

A quote-request funnel losing buyers at the final step, run through the loop

Illustrative engagement, not a client account

A B2B distributor’s site draws steady traffic to its product pages, but the quote-request form — the funnel’s only conversion — is completed by a small fraction of those who open it. The commercial team’s standing theory is that the form is too long; the design agency’s is that the page needs a refresh. Neither theory has ever been tested.

  1. 01Instrumentation first: field-level form analytics show most abandonment happens at one field — a required company registration number many visitors do not have to hand — not gradually across a long form.
  2. 02A falsifiable hypothesis replaces both standing theories: making that field optional at this stage, and collecting it later in the sales conversation, will raise completed requests without degrading lead quality.
  3. 03The test runs variant against control to its planned sample, with lead quality tracked through the sales team as the guard metric, so a win on volume cannot hide a loss on value.

Whatever the result, the organisation ends the cycle knowing something it can act on: either the fix ships with its effect measured, or a plausible theory is retired without a redesign being bought on its behalf. That is the product — decisions with evidence attached — and the redesign, if it ever happens, is commissioned against findings rather than fatigue.

Our role

What the experiment programme covers and which calls stay commercial

  1. 01

    A funnel you can read

    We instrument the steps, reconcile the numbers with your analytics, and leave a funnel view the commercial team checks weekly — because optimisation without trustworthy measurement is redecorating.

  2. 02

    Hypotheses from your visitors, in both languages

    Evidence gathering covers Arabic and English sessions alike, because a funnel serving Saudi Arabia often leaks differently in each language and an English-only analysis optimises half the audience.

  3. 03

    Tests run and reported honestly

    We design each experiment, run it to its planned sample, and report inconclusive as inconclusive — a programme that only ever announces winners is describing its analysis, not your funnel.

Pricing, offers and what the business is willing to promise are commercial decisions and stay yours; we supply the evidence on how the funnel responds to them, and the discipline that stops a lucky week being mistaken for a strategy.

Common questions

What buyers ask

What volume of traffic makes a conversion programme worth starting?+

Less than most teams assume, because split testing is only one instrument in the case. High-traffic funnels earn classic A/B tests. Lower-traffic funnels still support form analytics, session evidence, moderated usability sessions and the repair of outright defects — none of which need statistical volume, and all of which move conversion. The threshold question is not "do we have enough traffic to test" but "do we have enough conversions at stake to care" — and if the funnel carries real revenue or real leads, the answer is usually yes, with the method chosen to fit the volume.

We copied a competitor’s checkout design and conversion did not move. Why?+

Because you copied their answer without having their question. A design that wins on one funnel wins for reasons rooted in that business’s traffic mix, price points, brand trust and audience — none of which came with the layout. Their checkout may compensate for a weakness you do not have, or rely on a strength you do not have. This is the core argument for testing on your own funnel: the only visitors whose behaviour predicts your conversion rate are yours. Competitor funnels are a fine source of hypotheses; they are a poor source of conclusions.

How is conversion work different from just improving the design of the site?+

Design work optimises for coherence and brand; conversion work optimises for a measured behaviour, and the two sometimes disagree. A page can become more beautiful and convert worse. The practical differences are that conversion work starts from funnel data rather than a brief, changes one thing at a time where design changes many, and submits every change to a verdict — kept if it moved the metric, reverted if not. The disciplines are complementary: research and design build the experience worth testing, and conversion work is the accounting that verifies what the experience earns.

Start the conversation

A funnel audit that shows where visitors are actually leaving, and the first three testable hypotheses ranked by expected commercial value.

Start the conversation