I'm Hichem Bennaceur,
CRO lead for DTC brands.

I started as a software engineer and moved into marketing. Today I find where online stores lose sales, then write, design, build and test the fix myself.

Book a call See my results
Hichem Bennaceur
77experiments since April 2026
5DTC brands, 5 verticals
11live pages and funnels built from scratch

About me

My engineering background means a technical problem rarely blocks a test. I read the code, fix the tracking and ship the page myself, and I write copy from what customers say.

Today I lead CRO in-house for a supplement brand, where I took over from an agency, and work on experimentation programs for four other DTC brands.

Testing
Intelligems, Convert
Platforms
Shopify, Checkout Champ, Next Commerce
Measurement
Google Analytics, Google Tag Manager, order-platform reporting

Why work with me

Certified by top companies

CXL Certified Optimizer, plus product-led growth, Google Tag Manager and product analytics.

Clients on 4 continents

USA, UK, France, Malaysia and Algeria.

A marketer with deep technical knowledge

Software engineer turned marketer, so technical problems rarely block the work.

Track record across business types

SaaS, e-commerce and agencies, in analytics and CRO from A to Z.

Remote and async by default

5 years of fully remote work, with clear written updates.

Skills and certifications

CXL
ProductLed
Google Analytics
Semrush Academy
HubSpot
  • CRO
  • Web analytics
  • Technical SEO
  • Product analytics
  • Technical marketing
  • User experience (UX)
  • Go-to-market strategy
  • Marketing automation
  • Messaging and positioning

What I do

Conversion rate optimization

I find what stops people buying, then test the fix across pre-sell, product page, checkout and post-purchase.

Web analytics

I set up and repair tracking so your reports and test results can be trusted.

Product analytics

I track how people use your product, which features they keep, and where they get stuck.

Landing page optimization

I review copy, positioning, UX and clarity, then write, design and build the new page.

My results

Real tests from five DTC brands, across the whole funnel. Pick a step to see one. Brand names stay private.

Travel goods brand, pre-sell

A listicle between the ad and the product page

The problem. The ads with the best click-through converted worst. They told a story, then dropped people on a generic product page.

What I did. I split paid traffic. One arm went through a listicle that carried on the ad's story and handed off to the product page.

What came next. One pre-sell per dominant ad angle, because a new narrator is a new message-match problem.

MetricBeforeAfterChange
Conversion1.28%1.45%+13.8%
Order value$174.52$176.37+1.1%
Revenue per visitor$2.23$2.56+15.0%

Split test, 97.2% probability to beat control

Games brand, landing page

Five iterations on one subscription page

The problem. The page had a single purchase option, and scroll data showed visitors leaving 15 to 20% of the way down, before they reached any trust content.

What I did. I added a two-game bundle tier, moved the buy box into the first screen, cut the copy above the fold, and put proof and the guarantee at the decision point.

What came next. The two-game bundle later beat the three-game bundle, because three cost more per item.

MetricBeforeAfterChange
Revenue per session, iteration 1$0.71$1.98+178.9%
Revenue per visitor, iteration 5$0.71$2.34+230%

Before and after across five iterations

Home textiles brand, product page

True scarcity next to the buy box

The problem. The product ships in limited production batches, but the page never said so. Visitors had no reason to decide now.

What I did. I added a batch-availability widget wired to real inventory. No fake countdown.

What came next. The same widget lost on a second funnel, so it stays here. A winning pattern gets re-tested, never copied.

MetricBeforeAfterChange
Conversion2.18%2.57%+18.0%
Order value$198.61$209.39+5.4%
Revenue per visitor$4.32$5.38+24.4%

Split test, 94.9% probability to beat control

Wellness brand, product page

A product page rewritten in the customer's words

The problem. Ads spoke about the customer's problem. The product page led with materials and origin, so the story broke at the moment of decision.

What I did. Customer research ranked the problems buyers care about most. I rewrote the headline and bullets around that ranking.

What came next. Order value dipped, so the win came from more buyers. Next tests stay on conversion.

MetricBeforeAfterChange
Conversion6.80%7.39%+8.7%
Order value$111.02$109.27−1.6%
Revenue per visitor$7.55$8.08+7.0%

Split test, 99.7% probability to beat control

Supplement brand, checkout

Fix the measurement, then replace the checkout

The problem. A legacy checkout on the main funnel. The test tool was wrong too: a cookie mismatch undercounted control, and revenue tracking sent item price only.

What I did. I fixed tracking first and made the order platform the source of truth. Then I tested a cleaner checkout and bundle buybox as one branch, because the two are linked.

What came next. Ship the branch as tested, then tune the buybox tiers.

MetricBeforeAfterChange
Conversion1.68%1.92%+14.3%
Revenue per visitor$1.10$1.18+6.9%

Split test, about 50,000 visitors per arm, 99.8% on conversion

Supplement brand, post-purchase

Related products after checkout, instead of more of the same

The problem. The funnel only offered more units of what people had just bought, yet survey data showed 85% of them already took another supplement every day.

What I did. I rebuilt the post-purchase step around complementary products, launched as a plain first iteration with no video and no angle-specific copy.

What came next. Angle-matched copy and video, now that the offer itself is validated.

MetricBeforeAfterChange
Take rate27.5%34.6%+25.5%
Order value$51.68$57.60+11.5%
Revenue per visitor$14.23$19.90+39.9%

Split test, 99.8% confidence

Wellness brand, post-purchase

A post-purchase ladder with better decline paths

The problem. A declined upsell came back at the same discount as the downsell, and later slots clashed with a free-gift promo.

What I did. I kept the anchor offer, made the first downsell a better deal, added a higher-ticket item for engaged buyers and a low-priced one in the last slot.

What came next. Every slot gets judged on revenue per entering order, never on its own take rate.

MetricBeforeAfterChange
Downsell take rate4.63%10.09%
Revenue per order$9.88$11.55+16.9%

Before and after, funnel swap

What losing tests taught me

Price increase on every product

Kept control and moved to a $1 price ladder instead of one big jump.

A winning scarcity widget, on a second funnel

It looked like a win in week one, then reversed. This is why I don't call tests early.

How I work

  1. Find the constraint

    Surveys, reviews, analytics, scroll data and a page audit show where intent, trust or usability breaks.

  2. Rank the hypotheses

    Ideas get scored on impact, confidence and effort. One test owns each page zone, so results stay readable.

  3. Write and build it

    I write the copy, design the page and build it. Developers get a working build or a spec they can ship.

  4. Check tracking, then test

    Events verified, traffic sources separated. Revenue per visitor decides, with conversion and order value explaining why.

  5. Decide in writing

    95% is my bar, and anything below it gets labelled as such. Winners ship, losers get logged, and the next iteration is on the board.

Compared with a freelancer or agency

Working with meTypical freelancerTypical agency
Who sets strategy Me, with your goals You do Shared across teams
Who does the work Me, every time The person you hired Often a junior
Who you talk to One person, directly One person An account manager
Copy, design and build All three Usually one of the three Three people, three handoffs
Patterns from other clients Across DTC, SaaS and agencies Limited to a few clients Siloed by team
Speed to a decision Fast, no approval chain Medium Slow, several approval layers
What you pay for A predictable retainer Hours The whole team

Who I work with

DTC brands

Full-funnel CRO

Experiments across pre-sell, product page, checkout and post-purchase.

  • Hypothesis-driven A/B testing
  • Checkout and product page optimization
  • Clean analytics that make tests trustworthy
See DTC work

Agencies

White-label CRO

Analytics and experimentation at a fixed price, under your brand.

  • White-label by default
  • Fixed price, you set the margin
  • Agency-native workflow
See agency work

Early-stage SaaS

Signup-to-activation fixes

I map the journey from ad click to activation, then fix each point where you lose people.

  • Full AARRR funnel audit
  • Onboarding and activation CRO
  • Retention cohort analysis
See SaaS work

Private portfolio

Named brands, test readouts and the 11 live pages I wrote, designed and built. It needs an access code because it holds client data.

+230%revenue per visitor on one funnel, before and after five iterations

Open the portfolio No code? Request it and I'll send it.

What clients say

Rated 4.8 out of 5 by 26+ clients.

Book a call

30 minutes. Bring your funnel and any recent test results, and we'll see whether I can help.

Questions

How do you charge?
Hourly or by project, whichever suits the work. My hourly rate is $100 (USD), dropping to $90 for all future work once a project passes 50 hours. For project pricing, I send a quote within 48 hours of the audit and once the context is clear.
Which industries are a good fit?
Good fit: supplements and nutrition, wellness, home, travel goods, games, SaaS, software and tech, education, electronics, baby and pets, DIY, watches, cars.

Not a fit: poker and gambling, Web3, fintech and banking, food, fitness, cosmetics, women's fashion, tobacco, dropshipping.
Can I get an estimate first?
Yes. Book a call so I can understand the business and what I can add. If we agree it's a fit, you get an estimate within 48 hours. You can also email the context to behichemben@gmail.com.
Why are the brands anonymous?
These are live businesses and the numbers are commercial data. Named results, tool screenshots and the pages I built are in a private portfolio. Request the access code and I'll send it.
Do you work remotely?
Only remotely. I've worked that way for 5 years, on Slack, ClickUp, Jira, Trello and WhatsApp, with written updates so nobody has to wait for a meeting.