Conversion Work That Starts By Telling You Whether You Can Test At All

Research, prioritization, and testing where the traffic supports it. Where it does not, we say so and ship the changes instead of pretending a test would mean something.

Get My Free Audit

Komplex runs conversion rate optimization from Austin, Texas, with a presence in New York City and Reykjavik. We are a Google Partner and have run digital work for companies from local service businesses to global retail brands. The work covers research and audit, analytics review, hypothesis and prioritization, and A/B testing. Where a site is too small to test, we ship the changes directly instead. One senior lead owns the analysis, the decisions, and the reporting, and stays on the account.

Google Partner

Founded In 2014

You Own The Test Data

No Long Contracts

Most Sites We Are Asked To Test Cannot Support A Test

The arithmetic is unforgiving and nobody in this category likes saying it out loud. On a page converting in the low single digits, reading a modest change needs roughly a thousand to two thousand conversions in each version. Below that, the result means nothing. Not visits. Conversions. Plenty of businesses with healthy revenue do not produce that in a quarter. Run a test anyway and you get a number that looks like an answer and is really just noise wearing a confidence interval.

Isometric illustration of a marketer reviewing keyword research as Komplex ranks number one in search results.

So The First Thing We Do Is Work Out Which Side Of That Line You Are On

It takes an afternoon and it changes everything that follows. Above the line, testing is the right purchase and we build a programme around it. Below it, the answer is research, session review, and user testing with a handful of people. Then shipping bigger changes on judgement rather than shaving a button colour. That second version is less impressive to describe and it is what actually moves conversion for most companies we talk to. Anyone selling you a testing programme without asking about your conversion volume has not done the arithmetic either.

Ask Which Side You Are On

Conversion Rate Optimization Services, From The Research To The Read

Eight lines of work. The first three happen whether or not you can test, and on a smaller site they are most of the value.

Research And Heuristic Review

A structured walk through the conversion path looking for the faults that show up on almost every site. Cheap, fast, and it usually finds something worth fixing before any tool gets involved.

Analytics Review

What the data already says about where people leave, on which device, from which source. Most accounts have the answer sitting in them unread.

Session And Heatmap Review

Watching real people fail to do the thing you wanted. Uncomfortable and more useful than any survey, and it works at any traffic level.

Hypothesis And Prioritization

A ranked backlog with the reasoning attached, so the next change is chosen rather than picked. You get the document whether or not we run the tests.

A/B Testing

Design, build, run, and read. The sample size gets worked out before launch and the result is called at the end, not when it looks good.

Shipped Changes For Smaller Sites

Where testing cannot reach significance, we make the change and measure the before and after, with the limits of that stated plainly.

Ecommerce Conversion Work

Product page, cart, and checkout, where the leaks are largest and the order volume is often high enough to test properly.

Measurement And Reporting

Conversions tied to revenue where the tracking allows it, with the consent gaps named rather than smoothed over.

Start With The Free Audit

Send the URL and roughly how many conversions the site produces a month. That second number decides more than anything else you could tell us.

Get The Free Audit

WHAT YOU CAN ACTUALLY BUY

Four Conversion Volumes, And What Each One Can Really Buy

Find your row. This is the conversation most agencies have after you sign rather than before, and it is the single most useful thing on this page.

Monthly conversions	What a test could tell you	What we would do instead	How you would know it worked
Under a few hundred	Almost nothing. A test would need most of a year to reach significance, by which point the site has changed anyway	Research, session review, user testing with a handful of people, then ship the obvious fixes as one change	Enquiry volume over a longer window, and the qualitative reasons people were failing before
A few hundred to around a thousand	Only very large differences, and only on your highest traffic page. Small refinements stay invisible	One or two bold tests on the page that matters most, plus shipped changes everywhere else	A clear result on the big test, and a before and after you can defend on the rest
Around one to five thousand	Real testing becomes possible. Moderate changes can be read in a sensible window on your main templates	A running programme with a prioritized backlog, one test at a time, called on the numbers	Tests that reach significance, and a documented record of what won and what lost
Above five thousand	Most of what you would want to know, including smaller refinements and tests on secondary pages	A continuous programme, several tests in flight, with segments and device splits worth reading	A test velocity you can plan around and a compounding record across the site

Not Sure Which Row You Are In?

Most people underestimate it or count sessions instead of conversions. The free audit works it out from your analytics.

Find My Row

RANKED BY HOW OFTEN WE FIND IT

Six Places Conversions Leak, Ranked By How Often We Find Each One

These are not exotic. They are what turns up on site after site, in roughly this order, and most of them can be fixed without any testing at all.

1
The Form Is Too Long
Fourteen fields where five would do, because somebody in sales asked for details they could have collected on the call. Every field costs you people, and the ones at the bottom cost the most.
2
The Mobile Version Was Never Really Checked
It renders, so everyone assumes it works. Then you try to complete the form with a thumb on a moving train and discover the submit button sits under a sticky bar.
3
The Page Is Slow Where It Matters
Not the score, the feel. A page that takes a beat to respond when someone taps it reads as broken. Interaction to Next Paint became a Core Web Vital in March 2024 for exactly that reason.
4
The Offer Is Not Clear In The First Screen
Visitors cannot tell what you sell, who it is for, or what happens if they click. No amount of testing fixes a page that has not decided what it is saying.
5
There Is Nothing To Trust
No proof, no specifics, no reason to believe you rather than the other five tabs open. This is the most common fault on B2B sites and the least often mentioned.
6
Checkout Or Enquiry Has A Silent Failure
A step that errors on one browser, a form that returns a thank you and delivers nowhere, a payment method that fails quietly. Every form gets submitted during the audit, and the answer is often uncomfortable.

Want To Know Which Of These You Have?

The free audit checks all six from the outside and ranks them by what each is likely costing you.

Check My Site

What Is Conversion Rate Optimization?

Conversion rate optimization is the practice of increasing the share of visitors who do what a website was built for. That might be buying, enquiring, booking, or signing up. The work covers research into where people currently fail and hypotheses about why. Then prioritizing what to address, testing where traffic allows, and measuring the result. It is a research discipline before it is a testing one, which is why a good CRO engagement produces documents before it produces experiments.

A conversion rate optimization agency does that work on your behalf. It differs from web design, which builds the site, and from technical SEO, which makes it findable. Komplex runs this from Austin, Texas, alongside the paid and organic programmes that send the traffic. The page and the campaign feeding it get looked at by the same person.

WHO WE DO THIS FOR

Conversion Work For Ecommerce, B2B And Software, Home And Kitchen, And Local Services

The conversion path changes completely by sector, and so does whether testing is even available to you. Our accounts have covered footwear, skincare, software, telecom, and kitchens and home improvement.

See If We Fit

Ecommerce And DTC

Usually the only sector with enough order volume to test properly. Product page, cart, and checkout are where the money is, and our ecommerce conversion work goes deeper on it.

B2B And Software

Low volume, high value, and almost never testable. Here the work is form design, proof, pricing page clarity, and whether the leads are worth calling.

Home And Kitchen

Long decisions with a showroom visit in the middle. The site's job is to survive being compared, which is a different brief from closing.

Local And Professional Services

The conversion is often a phone call, which most analytics setups do not count properly. Fixing the measurement usually comes before fixing the page.

THE NUMBERS

What The Benchmark Data Says, And Why Two Sources Disagree

Four figures worth knowing before anyone quotes you a target. The first two look contradictory and both are correct.

Across 41,000 dedicated landing pages in one benchmark study, the median converted at 6.6 percent. Software sat near 3.8 percent and events above 12 percent.

A widely cited cross industry figure sits nearer 2.35 percent. It averages paid search landing pages fed by cold traffic, which is a different measurement.

Google Optimize shut down in September 2023 with no free replacement. Testing now carries a tool cost that did not exist before.

Interaction to Next Paint replaced First Input Delay as a Core Web Vital in March 2024, which made responsiveness a stricter measure than it had been.

HOW WE WORK

Look, Decide, Test Or Ship, Read, Then Do It Again

Five stages. The second one is where we tell you whether the third is testing or shipping, and that conversation happens before you commit to anything.

Look
Analytics, session recordings, the conversion path, and every form submitted for real. Plus the number that decides the rest, which is monthly conversions.
Decide
A ranked backlog with the reasoning attached, plus a plain verdict on whether your volume supports testing. You keep the document either way.
Test Or Ship
Above the threshold, we build and run tests with the sample size fixed in advance. Below it, we make the change and measure it within the limits that allows.
Read
Results called on the numbers rather than when they look good. A test that loses gets reported as a loss and tells you something worth knowing.
Do It Again
The backlog gets reordered from what we learned. Conversion work compounds, which is the whole reason it is a programme rather than a project.

See Where The Site Is Losing People

The audit comes first and there is no charge. It reads the path, the forms, the speed, and the volume, then ranks what to fix.

Start With The Free Audit

FREE CONVERSION AUDIT

Get A Free Conversion Audit

We walk the conversion path on desktop and mobile, submit every form, and check the speed on a real connection. Then we read what the analytics already says and work out whether your volume supports testing. The write up ranks what to fix by likely impact and tells you plainly which of it is worth paying anyone to test. If the verdict is fix three things and revisit next year, that is what the note will say. There is no charge, the findings are yours whatever you decide, and the same audit sits inside our free marketing audit.

Your conversion auditFree

The conversion path on desktop and mobile
Every form submitted for real
Form field count against what you need
Speed where it affects the feel
What your analytics already shows
Whether your volume supports testing

1
Tell us about the site
The URL, roughly how many conversions a month, and what you want more of.
2
We look
Read only access to analytics goes deepest. Nothing gets changed, and an NDA is available on request.
3
You get the scorecard and a call
A written scorecard ranked by impact, then a call to walk through it.

Ready To Find Out Where People Are Leaving?

There is no fee and no obligation, and everything the audit finds belongs to you.

Get My Free Audit

PRICING

How We Price Conversion Work

Komplex works on a monthly retainer, sized to the work rather than to a fixed package. We agree what is in scope before anything starts. We are open to other formats where one suits you better. Two things move the number. Whether the site can support a running test programme or needs research and shipped changes, and how many templates are in scope. A one page enquiry funnel and a store with forty templates are different jobs.

Testing tools are billed separately and bought in your name. Google Optimize shutting left a real cost, and it should sit with you rather than be marked up. We do not charge a share of any lift, and we would treat an agency offering to as a warning rather than a feature. Anything outside the agreed scope gets quoted before it starts.

The quickest route to a number is a short conversation about the site and the volume. Tell us both and we will explain how we would scope it and what we would leave out.

Get A Quote

What Clients Say About Working With Komplex

The Komplex team helped us reach new audiences and drive higher conversions.

Strategies that delivered results beyond our expectations.

Their expertise helped us scale efficiently, attract more customers, and drive measurable growth.

Komplex played a key role in growing our user base and optimizing our marketing efforts.

Common Questions About Conversion Rate Optimization

What is conversion rate optimization?

It is the practice of increasing the share of visitors who do what your site was built for. The work is research into where people fail, hypotheses about why, prioritizing what to address, testing changes where traffic allows, and measuring what happened. It is a research discipline before it is a testing one, and the documents usually arrive before any experiment does.

Do we have enough traffic to run a test?

It depends on conversions rather than visits, and the honest threshold is higher than most people expect. Reading a modest change needs roughly a thousand to two thousand conversions in each version. The table above shows what each volume band can actually buy, and the free audit works out which row you are in from your own analytics.

What is a good conversion rate?

It depends entirely on what is being measured. Dedicated landing pages measured across tens of thousands of examples show a median around six and a half percent. Software sits near four and events above twelve. A widely quoted cross industry figure near two and a half percent averages paid search pages fed by cold traffic. That is why the two numbers disagree without either being wrong.

Can you guarantee a lift?

No, and we would be careful with anyone who does. Too much sits outside an agency’s control: your traffic mix, your offer, your competitors, and the season. Several agencies sell guaranteed lift models, and several respected guides in this field treat a guaranteed percentage as a warning sign. We sell the method, the honesty about what it can prove, and the discipline to call a losing test a loss.

What do you do if we cannot test?

Research, session review, user testing with a small number of people, and shipping bigger changes on judgement. Then measuring the before and after and being clear about what that does and does not prove. For most companies this is the real answer, and on a smaller site it is where the improvement comes from.

How long before we see anything?

The research findings arrive early because they do not depend on volume. Test results depend entirely on your conversion volume, and we work out the honest window before anything launches rather than after. We do not commit to a date for a result that arithmetic controls.

How much does conversion work cost?

It is a monthly retainer sized to the work. What moves it is whether you can support a testing programme and how many templates are in scope. Testing tool licences sit with you rather than being marked up. The quickest way to a number is a short conversation about the site and its volume.

What is the difference between CRO and web design?

Web design builds the site. Conversion work improves a site that already exists, using evidence from how people are using it. If your problem is that the site is old, unusable, or cannot be edited, that is a web design job first and conversion work afterwards. We will say which you need rather than selling the one you asked for.

How does this relate to our landing pages?

A landing page built for a campaign is a different purchase from improving pages that already receive traffic. Building new pages sits with our landing page design work. This page owns testing and improving what exists. If you have no page yet, start there.

Who owns the test data?

You do. The analytics account, the testing platform account, the session recordings, and every document we produce sit in your name. We work through access you can remove at any time. The backlog and the test record are yours to take to anyone, and plenty of clients have.

What happens when a test loses?

We report it as a loss and say what we think it means. Losing tests are useful, because they stop you shipping something that would have cost money. A programme where everything wins has a measurement problem. Anyone showing you only winners is showing you a selection.

Do you work on Shopify?

Yes, and Shopify stores are often the best candidates for real testing because the order volume supports it. Checkout customization is limited by your plan, which is a constraint worth knowing before we plan anything. Our ecommerce conversion work covers the store funnel in detail.

Who does the work?

One senior lead owns the research, the hypotheses, the analysis, and the reporting, and stays on your account. Build work on tests goes to contract developers we have used for years, under that lead. This is detail heavy work, and quality drops fast when it gets passed to whoever has capacity.

Do you work with companies outside Austin, Texas?

Yes. Komplex is based in Austin with a presence in New York City and Reykjavik, and this work happens remotely by nature. Sites serving several markets get the conversion path checked per market, since a form that works in one country often fails on address format in another.

Send The URL And Your Monthly Conversion Count

Those two things decide almost everything. Back comes a ranked list of the places people leave, plus a verdict on whether paying anyone to test is justified.

Get My Free Audit

Research And Heuristic Review

Analytics Review

Session And Heatmap Review

Hypothesis And Prioritization

A/B Testing

Shipped Changes For Smaller Sites

Ecommerce Conversion Work

Measurement And Reporting

Ecommerce And DTC

B2B And Software

Home And Kitchen

Local And Professional Services

Across 41,000 dedicated landing pages in one benchmark study, the median converted at 6.6 percent. Software sat near 3.8 percent and events above 12 percent.

A widely cited cross industry figure sits nearer 2.35 percent. It averages paid search landing pages fed by cold traffic, which is a different measurement.

Google Optimize shut down in September 2023 with no free replacement. Testing now carries a tool cost that did not exist before.

Interaction to Next Paint replaced First Input Delay as a Core Web Vital in March 2024, which made responsiveness a stricter measure than it had been.

Look

Decide

Test Or Ship

Read

Do It Again

The Komplex team helped us reach new audiences and drive higher conversions.

Strategies that delivered results beyond our expectations.

Their expertise helped us scale efficiently, attract more customers, and drive measurable growth.

Komplex played a key role in growing our user base and optimizing our marketing efforts.

What is conversion rate optimization?

Do we have enough traffic to run a test?

What is a good conversion rate?

Can you guarantee a lift?

What do you do if we cannot test?

How long before we see anything?

How much does conversion work cost?

What is the difference between CRO and web design?

How does this relate to our landing pages?

Who owns the test data?

What happens when a test loses?

Do you work on Shopify?

Who does the work?

Do you work with companies outside Austin, Texas?

Paid Search Management

Landing Page Design

Web, App, And Design

Email Marketing

SEO Services

Display And Video Advertising