Harry Molyneux shares advice and case studies on conversion-rate optimization for direct-to-consumer ecommerce brands, including how to improve product pages, checkout, and A/B testing.
Harry Molyneux is useful for providing conversion rate optimization advice and case studies for direct-to-consumer e-commerce brands.
Best for
DTC supplement brands with high traffic seeking a conversion program for product pages, cart, and checkout.
Not for
Brands seeking traditional media buys like sponsored videos or paid newsletter mentions judged on impressions.
Key takeaways
Conversion work should be seen as a portfolio of experiments.
Product pages need better structure, not more content.
Retention is decided on the product page, not in post-purchase email.
Check first: Confirm LinkedIn follower demographics, job titles, and engagement figures for recent conversion-rate posts.
Main ideas
Run conversion work as a portfolio of experiments, and expect most to prove nothing · The traffic is already bought, so conversion work is the cheaper lever · Product pages need better structure, not more content
Harry Molyneux writes about conversion-rate optimization for direct-to-consumer ecommerce: how product pages, checkout, customer discovery, and testing affect sales.
Their posts pair specific changes with reported outcomes.
One example describes a beauty brand adding five ordered information sections below the add-to-cart button, with reported lifts in conversion rate and revenue per visitor.
Another explains the results of 1,379 tests, including the large share that were inconclusive.
Follow their work for practical ecommerce examples and a clear-eyed account of what testing can—and cannot—prove.
Molyneux connects store changes to their rationale and measurement.
They explain why the accordion’s information order might help shoppers, then report several performance measures.
They also treat inconclusive tests as important results rather than hiding them behind a win rate.
Their advice ranges from individual product-page changes to broader programs for discovery and repeat purchase, with occasional guidance on timing and test methods, such as using Meta rather than an underpowered on-site A/B test for low-traffic Black Friday pages.
The posts also cover agency hiring and operations, a smaller, internal-facing part of the work.
Harry Molyneux writes about conversion-rate optimization for direct-to-consumer ecommerce, mostly from the position that conversion work should be a portfolio of measured experiments rather than a redesign.
Read more
The recurring ideas are a blunt base rate for testing, an argument that the traffic a brand already paid for is the cheapest revenue available, a preference for rearranging existing product-page content over adding more of it, and a view that subscription retention is decided at the point of sale.
A second strand covers running the agency itself, where Molyneux frames quality problems as hiring problems.
The figures throughout are the author's own reported numbers from client work, and no publication dates are given in the posts here.
Run conversion work as a portfolio of experiments, and expect most to prove nothing
This is the position underneath everything else Molyneux publishes.
Read more
A test only counts as a win when it beats the control group with statistical significance on a KPI treated as relevant for that test — conversion rate, revenue per visitor, average order value, products per visitor.
If the significant movement goes the other way it is a loss, and anything in between is called inconclusive rather than reported as a partial win.
The point Molyneux draws is deliberately unflattering: the interesting number is not how often testing works but how often it does not, so store changes made on a hunch are usually doing nothing and occasionally costing money.
Across 1,379 tests run for more than 40 clients over five years, 523 won (37.9 percent), 325 lost (23.6 percent) and 531 were inconclusive (38.5 percent). Molyneux reads that split as evidence that two thirds of tests fail to prove they made the client money.
The traffic is already bought, so conversion work is the cheaper lever
Molyneux argues from an asymmetry in how brands treat spending.
Read more
Founders commit large quarterly budgets to paid social without asking for a guarantee, then want a guarantee before approving conversion work.
The reasoning offered is that every visitor on the site has already been paid for, so converting them costs no additional media spend, and the real question is how much revenue is being lost each month by guessing at what converts.
It is an argument designed to be reused inside a brand rather than a description of a technique.
Molyneux reports a 4-to-8x average return within 90 days, tracked across clients for two years; a brand spending around $10M a month that ran more than 140 experiments; a checkout test credited with $156,000 in extra monthly revenue; and product-page review placement credited with $53,700 a month.
Product pages need better structure, not more content
The recurring complaint is that DTC brands respond to weak product pages by adding — more hero images, more testimonials, more trust badges.
Read more
Molyneux argues shoppers are already hunting for specific answers and skipping past that material, so the lever is grouping and ordering the answers they want, not producing more of them. The proposed order differentiates before price is shown, gives concrete detail once value is understood, and closes with logistics and reassurance.
A beauty brand added an accordion below the Add to Cart button with five sections in a fixed order: product description, how the brand is different, what is included, how it works, and shipping and returns.
Reported results: +2.74 percent conversion rate, +5 percent revenue per visitor, +2.19 percent average order value and +4.20 percent products per visitor, or about $19,705 in added monthly revenue, with the largest lift among new users.
Two posts describe the same case in near-identical wording; count it as one result.
Retention is decided on the product page, not in post-purchase email
Molyneux's argument is that most subscription cancellations happen in the first month and are caused by vague expectations rather than the product: the customer does not know what the product working should feel like or when to expect it, so they conclude nothing is happening and cancel.
Read more
The fix proposed is to define the success timeline at the point of sale, so the buyer has something specific to look for and by the end of the first month the habit is formed. This places churn prevention in front of the buy button rather than in lifecycle messaging.
A close reading of AG1’s product page describes a module called Your 30-Day AG1 Journey, staged as energy support for days 1 to 7, immune support for days 8 to 14, and digestion for days 15 to 30.
Supplements face a trust problem about efficacy, not safety
This is Molyneux's most category-specific argument: a supplement brand is asking someone to believe a product will work, on a phone, with no physical proof, so the barrier is legitimacy and efficacy rather than whether the product is safe.
Read more
From that premise the recommended work is social proof placed at each decision point, subscription messaging that does not cost conversion, mobile friction removal, short product-discovery quizzes, and trust signals positioned where checkout doubt peaks. The wider claim is that the work is removing invisible barriers for buyers who are already convinced rather than rebuilding the store.
Molyneux reports 11 social proof tests for one brand that all won, with reviews naming specific pain points outperforming generic testimonials, and cites over 80 percent of supplement traffic arriving on mobile, where simple fixes such as a back-to-top button are claimed to add six figures monthly.
Buy the customer before the discount, and do not test pages you cannot power
Molyneux sets seasonal work months ahead on a cost argument: because every brand floods paid social at the same time, acquisition costs multiply in the sale period, so subscribers acquired early are far cheaper than the same people bought on the day.
Read more
The sequence is to acquire subscribers first, lock the offer before any page is designed, and build one landing page per offer and per ad angle so cold traffic does not land on a homepage. The testing advice is where this sits alongside the strong advocacy of testing elsewhere: below roughly 100,000 monthly visitors a traditional on-site test of those pages will not reach sample size, so winners get called from noise and the pages should be tested in the ad platform instead.
The post lays out a 13-week countdown to Black Friday: run paid campaigns optimised for email signups rather than purchases while CPMs are around $2 instead of $8 or more, then stop a losing bundle test at nine days and launch the next idea the same day. It also says most DTC brands make 30 to 50 percent of annual revenue in those roughly 60 days.
Agency quality is a hiring problem before it is a growth problem
Away from client tactics, Molyneux writes about running the business, and the consistent argument is that operational chaos is a staffing failure rather than a strategy failure.
Read more
One post describes remote hiring where candidates hold several full-time contracts at once and underperform everywhere, leading to a process change of contacting previous managers directly instead of accepting references from candidates. Another describes a year of winning clients, letting quality slip and losing clients, and attributes it to growth outpacing the ability to hire and train rather than to positioning or service design.
The stated fixes after almost 12 months of that client-churn cycle were a full-time head of people dedicated to recruiting, paid technical assessments before anyone joins, immediate handling of underperformance, and faster decisions to let people go. Molyneux claims retention improved afterwards and summarises it as a chaotic agency having a talent problem rather than a growth problem.
If you are deciding whether to fund a testing program, start here: it answers the base-rate question most proposals dodge—how often testing produces a provable result.
Read more
Harry Molyneux also states the measurement rules plainly, so you can compare them with your own.
Across 1,379 tests run for more than 40 clients over five years, 523 won (37.9 percent), 325 lost (23.6 percent) and 531 were inconclusive (38.5 percent).
A test counts as a win only when it beats the control on a relevant KPI—conversion rate, revenue per visitor, average order value or products per visitor—with statistical significance.
The takeaway: two thirds of tests do not move a proven metric.
Changes made on a hunch are guesses, and some cost money.
For brands planning a campaign
Budget for test volume and time, not a lift from every idea. Agree upfront on what counts as a win, how to handle inconclusive tests and who decides when a result is noise.
For a supplement or wellness campaign, it names a specific barrier and gives a prioritized list rather than general advice.
Supplement brands face a trust problem around efficacy and legitimacy: shoppers have no physical proof that a product will work when they see it on a phone.
Molyneux says the team has worked with seven-, eight- and nine-figure supplement brands and run at least 100 tests on them.
Six practices follow: add social proof at each decision point (11 social proof tests for one brand all won, with reviews tied to specific pain points beating generic testimonials); move buyers to subscriptions without hurting conversion; prioritize mobile fixes, given that over 80 percent of supplement traffic is mobile (a back-to-top button is claimed to add six figures monthly); keep quiz funnels short for multi-product brands; place trust signals where checkout anxiety peaks, at card entry; and use long-form education while keeping a path back to the buy box.
For brands planning a campaign
For a supplement store, prioritize social proof and mobile friction before a redesign. Keep quiz funnels short enough to protect completions.
This is the clearest before-and-after in the material, with the change described precisely enough to reproduce.
Read more
It is a strong starting point if you want one concrete product-page idea rather than a philosophy of testing.
A beauty brand added an accordion directly below the add-to-cart button, with five sections in a fixed order: product description, how the brand is different, what is included, how it works, and shipping and returns.
The logic: show differentiation before price, add tangible detail once value is clear, then close with reassurance.
Reported results were conversion rate up 2.74 percent, revenue per visitor up 5 percent, average order value up 2.19 percent, products per visitor up 4.20 percent and about 19,705 dollars in added monthly revenue.
The lesson is structure, not volume: users are hunting for answers and skipping past hero images.
A later post repeats the case with near-identical wording and numbers, so count it as one result, not two.
For brands planning a campaign
Prototype the accordion structure on a live product page and measure it against the same KPIs. It is a low-build change that tests information architecture without producing new content.
This shows how Harry Molyneux analyzes another brand's page and turns it into a lesson—a useful way to assess their thinking before hiring.
Read more
It also spells out where they believe churn is decided.
AG1's product page features a module called Your 30-Day AG1 Journey, with days 1 to 7 for energy support, 8 to 14 for immune support and 15 to 30 for digestion.
The analysis: what looks like standard benefit copy also serves as retention psychology.
Most subscription cancellations happen in month one and are driven by vague expectations rather than the product itself.
Defining the success timeline at the point of sale gives customers something specific to look for; by day 30, the habit is formed.
Retention starts on the product page, not in email flows.
For brands planning a campaign
For subscription brands, test an expectation-setting module on the product page and measure first-month cancellation, not just initial conversion. The same approach can work for non-subscription products whose results take weeks.
The accordion post shows one change; this one shows how the work can span a whole store.
Read more
It is a useful example for judging whether they think in programs or isolated tests.
The client is described as already performing well, with a healthy conversion rate, loyal customers and a strong product lineup.
Three workstreams follow: restructuring navigation and search for product discovery, reported as reaching six figures in monthly revenue; repurchase incentives, described as two of the agency's biggest test wins, with existing customers buying more often; and brand-first positioning, chosen after surveys showed recognition mattered most.
Reported totals: 13 winning tests, roughly 1 million dollars in new revenue within months and 13 million dollars projected annually.
The summary: good sites convert and great sites scale.
For brands planning a campaign
Scope a similar engagement around discovery, repeat purchase and positioning. Separate measured revenue from projections when presenting the plan internally.
Open this if you already believe testing matters and need to convince someone else.
Read more
It frames the spending decision in terms a media buyer or finance lead already accepts.
The argument starts with an asymmetry: founders commit large quarterly ad budgets to Meta without demanding a guarantee, then ask for guarantees on a conversion optimization proposal.
The traffic is already paid for, so converting it requires no extra media spend; the real question is how much is lost each month by guessing.
Supporting figures: an average return of 4 to 8 times within 90 days, tracked across clients for two years; a brand spending 10 million dollars a month that ran over 140 experiments; a checkout test credited with 156,000 dollars in extra monthly revenue; and a product-page review-placement test with 53,700 dollars monthly.
For brands planning a campaign
Frame the proposal as recovering value from traffic you have already bought, and ask for test budget rather than media budget. Treat the return range as the team's reported track record.
A rare calendar post, and the most directly usable item for a brief with a Q4 deadline.
Read more
It also warns against testing on low-traffic sites.
The sequence starts 13 weeks before Black Friday: acquire email subscribers with Meta campaigns optimized for signups, not purchases.
CPMs triple or quadruple in November, the post argues, making early subscribers a fraction of the cost of acquiring them on the day.
Lock the offer before design work, then build a landing page for each offer and ad angle; cold traffic from a specific ad should not land on the homepage.
For sites with fewer than 100,000 monthly visitors, traditional A/B tests will not reach sample size, the post warns, and risk producing a winner based on noise.
Test landing pages on Meta instead.
Kill losers quickly: one bundle test was stopped after nine days, and the next idea launched the same day.
The post also claims 30 to 50 percent of annual revenue falls in these 60 days, and that starting in mid-November is already too late to acquire subscribers efficiently.
For brands planning a campaign
Work backwards from the sale date and choose your testing method based on traffic: on-site tests above the traffic threshold, media-level tests below it.
Read this before relying on the case studies: it is where the author describes the agency's operating constraints.
Read more
That matters when assessing whether a small agency can absorb your account.
The post describes almost 12 months of a cycle: winning a client, delivering results, winning another, letting quality slip and losing a client.
The stated diagnosis is that growth outpaced the agency's ability to hire and train good people—not strategy or positioning.
Changes included hiring a full-time head of people dedicated to recruiting, requiring paid technical assessments before anyone joins, addressing underperformance immediately and firing faster.
The post calls this the kindest option for the team and clients, and claims retention improved afterwards.
Its summary: a chaotic agency has a talent problem, not a growth problem.
For brands planning a campaign
Ask who will run the work and how continuity will be maintained if staff change. Treat the retention claim as the author's account, not a verified metric.
Harry Molyneux publishes testing data, not just opinions.
Read more
One post reports results across 1,379 A/B tests run for more than 40 clients over five years: 523 wins (37.9%), 325 losses (23.6%) and 531 inconclusive (38.5%).
Their reading is that two-thirds of tests fail to prove a gain, so changes made on a hunch are usually neutral and occasionally costly.
Another post reports track-record totals of 4–8x return within 90 days across clients over two years, alongside single-test examples: one checkout test worth $156,000 in extra monthly revenue and product-page review placement worth $53,700 a month.
A test counts as a winner only when it delivers a statistically significant increase in a KPI deemed relevant for that test: conversion rate, revenue per visitor, average order value or products per visitor.
An increase in the other direction makes it a loser.
Anything between the two is inconclusive, not a partial win.
For brands planning a campaign
Brief this as a test roadmap with named success metrics and realistic expectations, not a one-off redesign.
Budget for a portfolio of experiments: roughly a third are expected to produce provable gains.
Before each test starts, ask which KPIs will determine its outcome.
A post sets out a supplement-specific playbook, arguing that the trust barrier is efficacy and legitimacy rather than safety: buyers have no physical proof that the product works.
Read more
It offers six practices: place social proof at every decision point, noting that 11 social proof tests for one brand all won and reviews naming specific pain points beat generic testimonials; move buyers to subscriptions without losing conversion through clear value and transparent pricing; prioritise mobile fixes, since more than 80% of supplement traffic is on mobile, with simple changes such as a back-to-top button credited with six-figure monthly gains; use short quiz funnels for multi-product brands, since every extra question costs completions; address checkout anxiety where buyers enter their card details, including with reviews that match cart contents; and provide long-form education with an easy route back to the buy box.
For brands planning a campaign
For a supplement or nutrition brand, use these six areas to set the agenda for a scoped engagement: social proof placement, subscription messaging, mobile friction, product discovery and checkout reassurance. Expect recommendations to focus on removing barriers for already-convinced buyers rather than rebuilding the store.
Scaling a healthy store by testing friction points
A case post describes a sports-nutrition brand with a healthy conversion rate, loyal customers and a strong product line.
Read more
Instead of a redesign, the work targeted specific friction points: restructuring navigation and search, credited with six figures in monthly revenue; adding repurchase incentives, described as two of the agency's biggest test wins because existing customers bought more often; and prioritising brand-first positioning after surveys showed recognition mattered most to buyers. Reported totals: 13 winning tests, about $1M in new revenue within months and a $13M projected yearly lift.
For brands planning a campaign
This suggests an approach built around incremental improvements to a functioning store, not a rescue.
If a brand already converts reasonably well, frame the brief around compounding revenue from known friction points.
First, ask for a prioritised list of those points.
A beauty brand client added a five-section accordion directly below the Add to Cart button and gained nearly $19,705 in monthly revenue.
Read more
The order was deliberate: product description, how the brand is different, what is included, how it works, then shipping and returns — differentiating before price, adding detail once value is clear, and closing with reassurance.
Reported results: +2.74% conversion rate, +5% revenue per visitor, +2.19% average order value and +4.20% products per visitor, with the biggest lift among new users.
The stated insight: shoppers hunt for answers and skip past hero images and paragraphs, so pages need easier access to existing content, not more content.
Treat product-page information architecture — grouping and ordering the answers shoppers look for — as the lever, ahead of adding imagery, testimonials or trust badges.
Two posts describe the same accordion result in near-identical wording. Treat them as one case with one set of numbers, not two independent wins.
For brands planning a campaign
A low-cost, concrete starting test for a DTC brief: reorder existing product-page content into a compact accordion below the buy box, then measure conversion rate, revenue per visitor and average order value.
Address retention on the product page, before the first order ships
A close reading of AG1's product page describes a 30-day journey module that sets expectations in stages: energy support for days 1–7, immune support for days 8–14, digestion for days 15–30.
Read more
Molyneux argues that most subscription cancellations happen in the first month because expectations are vague, not because of product quality: customers do not know what working feels like or when to expect it, so they cancel around week two. Defining the success timeline at the point of sale puts retention psychology in front of the buy button rather than in post-purchase email.
This is an observation of another company's page, not a result from one of Molyneux's own clients. Ask for client-specific evidence before treating it as a measured lift.
For brands planning a campaign
For a subscription brand, add expectation-setting content to the product page — what customers should notice and when — and judge it on first-month retention, not just conversion.
Seasonal planning that acquires before discounting
A Black Friday post lays out an order of operations 13 weeks ahead of the date: acquire email subscribers now, while Meta CPMs are around $2 rather than the $8 or more seen on the day; lock the offer before designing any pages; build one landing page per offer and ad angle so cold traffic does not land on a homepage; and kill losing variants quickly.
Read more
One bundle test was stopped at nine days and replaced the same day. The post notes that most DTC brands make 30–50% of annual revenue in those roughly 60 days, and brands starting in mid-November have already lost the cost-efficiency argument.
For brands with under 100,000 monthly visitors, Molyneux advises against traditional A/B tests on seasonal landing pages: the sample size will not be reached, and any winners will be called from noise. Instead, test the pages directly on Meta.
For brands planning a campaign
Seasonal work is planned weeks ahead and competes for attention early. Ask about availability well before the Q4 window, and tie the brief to subscriber acquisition and offer architecture, not just discount design.
Two posts shift from conversion tactics to running the agency.
Read more
One describes remote hiring problems, including candidates holding two or three full-time contracts at once, and a move to contacting previous managers directly rather than relying on candidate-provided references.
The other describes a year when growth outpaced the ability to hire and train, creating a cycle of winning a client, seeing quality slip and losing a client.
The response was to hire a dedicated head of people, introduce paid technical assessments before joining, address underperformance immediately and make faster decisions to let people go.
For brands planning a campaign
This is relevant to capacity planning: staffing and quality control are live operational concerns. Ask who will run the work and how senior attention will be protected during busy periods.
Harry Molyneux's posts show them working as part of a CRO agency team serving DTC ecommerce brands, including unnamed supplement, beauty and sports-nutrition clients.
Read more
Ask for the client names to assess their category experience. The posts name AG1, but disclose no sponsorship or partnership.
AG1
Mention
Harry Molyneux analyzed AG1's product page and its "Your 30-Day AG1 Journey" module as a retention lesson, calling it a masterclass from the biggest supplement brand.
Read more
The post doesn't disclose a sponsorship, client or partnership; ask whether the mention involved one.
Harry Molyneux publishes LinkedIn posts about conversion-rate optimization for direct-to-consumer ecommerce.
Read more
Start with the post reporting a 37.9% win rate across 1,379 A/B tests, which lays out the measurement-first approach behind the rest of the work.
The supplement CRO playbook covers social proof, subscription messaging, mobile friction and product quizzes; the beauty-brand accordion case examines product-page structure; the AG1 close reading focuses on retention on the product page; and the Black Friday countdown covers seasonal planning.
Another strand covers agency hiring and quality control.
For similar ecommerce conversion work, compare Jitendra kumar's introductory CRO and lead-generation checks, Hayan Khan's product-page redesign case work, and Md Moktadir Hossain's Shopify landing-page structure and trust signals.
Both publish LinkedIn posts about conversion-rate optimization, showing how to turn existing website traffic into more sales.
Read more
Jitendra kumar covers what CRO is, where to start on home, product, cart and checkout pages, and what to check when a website gets traffic but few leads.
Jitendra kumar writes introductory, checklist-style advice for websites and lead generation across service businesses. Molyneux reports measured A/B test results, focusing on DTC Shopify stores, supplements and subscription retention.
Both analyze ecommerce product pages to explain why shoppers convert or do not, covering benefit-first layout, purchase barriers, reasons to believe and the order of a buying journey.
Read more
Hayan Khan writes as a CRO designer, using a specific redesign case with conversion-rate and revenue before-and-after figures, including a critique of AI-generated design. Molyneux's work centers on testing base rates, supplements, subscriptions and seasonal planning.
Both write about Shopify landing pages and conversion optimization, covering clear headlines, trust signals, mobile-first design and why traffic alone does not produce sales.
Read more
Md Moktadir Hossain's posts offer general landing-page checklists and services for Shopify store owners. Molyneux's posts are more specific, drawing on client test data, supplement-category recommendations and Black Friday acquisition planning.
For a campaign aimed at DTC operators, the practical question is who sees Harry Molyneux’s work.
Read more
The posts give one measured audience figure: a LinkedIn following of 6,823, observed 9 October 2026.
That is a modest channel—enough to signal topical credibility with DTC operators, but not to carry a campaign as a reach play on its own.
Treat LinkedIn as a demand-generation and authority channel that supports a paid engagement rather than replacing it.
The posts don’t identify who follows them.
Their content targets people running or working inside DTC and Shopify stores: it speaks directly to founders, cites work with 7-, 8- and 9-figure supplement and sports nutrition brands, and closes with calls to book a strategy call.
A smaller share addresses agency operators on hiring and retention.
The intended reader is an ecommerce operator or brand owner, not a general consumer audience.
Before committing, ask for LinkedIn follower demographics or a follower survey covering job titles, seniority, company size and geography, plus engagement figures for recent conversion-rate posts.
Platform demographics can cover age, gender and geography, but employer and seniority require survey data or another documented source.
Harry Molyneux's LinkedIn posts describe client conversion-optimization work for DTC brands rather than sponsored placements.
Read more
They report experiment and case-study results—a product-page accordion that added about $19,705 a month for a beauty client, checkout and social-proof tests, and a Black Friday funnel sequence built on email acquisition before November—and close with invitations to book a strategy call or apply for agency help.
For buyers, this points to a CRO consulting or agency engagement, not a media buy.
There are no published sponsored-content formats, rates, deliverable specifications, exclusivity terms, or partnership inquiry route or contact link, so ask directly about scope, deliverables, and commercial terms.
Brief A (strong fit): a DTC supplement brand large enough to spend meaningfully on paid media — say around 200,000 monthly sessions, with a subscription option and a product page that has accumulated more content than structure — wants more revenue from traffic it already buys and asks for a conversion programme covering product page, cart and checkout.
Read more
This is the closest match in Harry Molyneux’s posts, and the match is specific.
They write as a conversion practitioner for DTC stores and have published a supplement-specific playbook: supplements face a trust problem about efficacy and legitimacy rather than safety, because the buyer gets no physical proof from a phone screen.
The six practices map directly onto this campaign: social proof at every decision point, subscription messaging without a conversion penalty, mobile friction fixes, short quiz funnels, trust signals placed at card entry where checkout anxiety peaks, and long-form education that keeps an easy route back to the buy box.
The post cites concrete results: eleven social proof tests for one client, all winners, with reviews naming specific pain points beating generic testimonials, and more than 80 percent of supplement traffic on mobile.
The product-page ideas are reproducible, which matters for a campaign asking for specific PDP work.
The clearest is an accordion directly below the add-to-cart button, with five sections in a fixed order — what it is, how the brand is different, what is included, how it works, then shipping and returns — so differentiation lands before price and reassurance closes.
Reported effect: conversion rate up 2.74 percent, revenue per visitor up 5 percent, average order value up 2.19 percent, products per visitor up 4.20 percent, and about 19,705 dollars in added monthly revenue.
The two near-identical posts describing it report one result, not two.
For a subscription brand, the second idea is an expectation-setting module on the product page, modelled on a reading of AG1’s 30-day journey layout: state what the customer should notice in days 1–7, 8–14 and 15–30 to address first-month churn before the first order ships.
The stated approach also fits the campaign’s A/B testing requirement.
Across 1,379 tests for more than 40 clients over five years, 37.9 percent won, 23.6 percent lost and 38.5 percent were inconclusive.
A win means a statistically significant gain on a KPI chosen for that test — conversion rate, revenue per visitor, average order value or products per visitor.
Molyneux’s reading is that two thirds of tests fail to prove a gain, a realistic frame for a programme budget.
Treat this as a paid CRO engagement, not a creator media buy: the posts describe consulting and agency engagements, not sponsored placements.
Before signing, ask for references, the measurement method behind the headline numbers, and confirmation of scope, fees and availability.
Reach is not the asset being bought; the one measured audience figure is a LinkedIn following of 6,823, observed 9 October 2026.
When to choose a specialist
Brief B (poor fit): a skincare brand wants three sponsored videos plus a paid newsletter mention across 90 days, bought on reach and judged on impressions and cost per thousand views.
Read more
Three things in Harry Molyneux’s posts argue against this skincare sponsorship campaign.
First, they describe no sponsorship offer to buy: no sponsored-content formats, rates, deliverable specifications, exclusivity terms or brand inquiry route.
The commercial pattern is an invitation to book a strategy call or apply for agency help, pointing to consulting rather than a media buy.
Second, the one measured audience number is a LinkedIn following of 6,823, observed 9 October 2026.
No engagement rate, follower demographics or geography is stated, leaving little basis to assess a reach-based campaign.
Third, the publishing itself works against sponsored video.
The strongest posts are analytical close readings: a walk-through of how a client’s product page was restructured and what changed, or how another company’s page handles churn.
That is teaching material aimed at operators, not personal-following content that carries a product to an audience; wrapping it as an ad would undercut its value.
A newly launched brand with around 15,000 monthly sessions wanting an on-site A/B testing programme across its store is also a poor fit as written.
Molyneux’s guidance is that below 100,000 monthly visitors a traditional test will not reach sample size and someone will call a winner from noise; the recommendation is to test landing pages on Meta instead.
The campaign would need to shift from site experimentation to media-level testing, with different deliverables.
That is a scope mismatch, not a reason to rule out an engagement: the same guidance would reshape a low-traffic Q4 plan rather than reject it.
A worked campaign example
Worked campaign angle for Brief A — a 90-day conversion programme for a supplement brand with a subscription option.
Frame it as a test roadmap, not a redesign.
The engagement produces a prioritised set of experiments on the product page, cart and checkout, each run against a control, with a written decision rule agreed before the first test starts.
Agree upfront that a win means a statistically significant gain on a KPI named for that test — conversion rate, revenue per visitor, average order value or products per visitor — and report everything else as inconclusive, not as a partial win.
Also agree who decides when a result is noise.
Set a kill rule: if a variant is down across the board, stop it within days; one bundle test was stopped at nine days and replaced the same day.
Budget on the expectation that roughly a third of tests produce provable gains.
Weeks 1–3, product page.
Prototype the accordion below add-to-cart with the five sections in the order above, using content the brand already has, and measure conversion rate, revenue per visitor and average order value.
If the store sells subscriptions, add a second variant with an expectation module on the same page — what the customer should notice in week one, week two and the rest of the first month — and judge it on first-month retention as well as initial conversion.
The rationale is that most cancellations happen in month one because of vague expectations rather than product quality.
Weeks 4–6, cart and checkout. Move trust signals to the point of card entry and test reviews matched to cart contents, on the reasoning that doubt peaks at payment rather than at the top of the page.
Weeks 7–9, mobile. Audit the mobile path for friction first: more than 80 percent of supplement traffic is mobile, and small mechanical fixes are the cheapest change available.
Weeks 10–12, discovery and repeat purchase.
If the store carries multiple products, test a short quiz funnel, keeping the question count low because every extra question costs completions.
For repeat purchase, test an incentive aimed at existing customers; in one published programme, this produced two of the largest test wins.
If the calendar includes Q4, move subscriber acquisition ahead of the discount season, buying email signups at normal CPMs rather than competing for the same person in November.
Report a smaller set of findings with more weight behind each, rather than every micro-variant.
Separate measured results from projections: the public material contains both — about a million dollars of new revenue within months against a 13 million dollar projected annual lift on one client.
Before signing, ask who will run the work day to day and how continuity will be handled if staff change; Harry Molyneux writes openly about hiring, training and retention.
Choose Jitendra Kumar for a first conversion audit of a site that is not a DTC store — a service business or general website — when the team needs the vocabulary of conversion work before commissioning tests.
Read more
The two posts offer a plain-language primer and a first-pass checklist.
The primer defines conversion rate optimization with a worked example: 1,000 visitors a day and 10 sales is 1 percent, and doubling the rate doubles sales without buying more traffic.
It names the pages where people drop off (home, product, cart and checkout), lists what can be optimized (product descriptions, button text and colour, page speed, images, trust signals), explains A/B testing, stresses mobile and frames the work as continuous rather than one-off.
The second post offers a six-question intake for a site that gets traffic but few leads: is the message clear, is the call to action findable, does the site build trust, is there too much friction, does it work on mobile, and does the page match the intent of the click.
For a brand that has never run a test, both are usable starting points that cost nothing to act on.
The advice stays at the level of what to check, with no reported test results, win rates or client examples.
Neither post covers supplements, subscription products or DTC checkout behaviour.
There is also no published programme for delivering a 90-day roadmap: no test design, sample-size discussion or decision rule for calling a result.
Where Harry Molyneux's material supplies measured outcomes and a stated rule for what counts as a win, this comparison offers general good practice that a brand would still have to turn into its own experiment plan.
It is better for orientation, but a poor fit for a brief that must be proved with controlled tests.
Choose Hayan Khan when the deliverable is a page rather than a programme: a single product page or a store navigation system designed or rebuilt around the shopper's decision.
Read more
The supplement redesign example offers a reusable method: lead with the benefit instead of a broad brand statement; bring the three purchase barriers of convenience, taste and cost onto the page instead of making shoppers hunt for them; add reasons to believe (ingredients, benefits, testing and certification, nutrition, functional claims, social proof); and follow a sequence of problem, solution, benefits, proof, product, purchase so every section has a purpose.
Reported movement in that example runs from 1.8 percent to 5.8 percent conversion rate and from 18,000 to 58,000 dollars in revenue.
The navigation post takes a different angle on the same buying decision: organise categories around problems customers already recognise — allergy and itchiness, joint health, stress and anxiety, gut health — rather than the brand's product names; add a quiz for shoppers who still need guidance; and start with the language customers use in reviews and support messages.
The post presents this as a hypothesis to test, not a proven cause.
The tradeoff is scope and evidence.
Both posts describe page-design decisions rather than a repeatable testing operation.
One redesign is a before-and-after, not a series of controlled experiments, and its impact story is incomplete: the redesign post says average revenue per user is not reported and some figures still need filling in.
The navigation example explicitly says it does not prove the menu drove the company's growth.
A brand seeking a prioritised product-page, cart and checkout test roadmap still has to build the measurement side itself.
Neither post offers a stated win rate or a rule for stopping losing variants early.
Choose this alternative when the goal is a better page; it is a weaker fit when results must be demonstrated through ongoing controlled tests.
Choose Md Moktadir Hossain when the problem is a dedicated landing page carrying paid traffic and the need is design and build, not experimentation.
Read more
Both posts start from the same diagnosis — traffic is not the problem, the page is — and list the same failure points: too many distractions such as popups and sliders, an unclear headline, no trust signals, a confusing layout, a missing value proposition and weak product presentation.
The remedy is a checklist a designer can execute: a clean, focused hero; strong product storytelling; social proof in the form of real reviews and photographs; one clear call to action rather than five different buttons; fast loading; and a mobile-first layout.
The test is whether a visitor can tell in five seconds what is being sold, why the brand can be trusted and why to buy now.
This is a fair fit when an ad and its landing page do not match, or when a page needs to be built around a single click rather than a whole-store audit.
The guidance is observational and unquantified.
Neither post reports test results, lift figures, client examples or sample sizes.
The closing calls to action offer to review a page on request, leaving no measured results to compare with a budget or target.
There is no framework for calling a winner and no treatment of cart or checkout, which matters when testing changes to those stages.
The value is design direction and a fast checklist, not a tested method.
Choose this alternative for one better page; it is insufficient when results must be proved with a controlled test.
Ten LinkedIn posts attributed to Harry Molyneux cover ecommerce conversion optimization, A/B testing, product-page structure, subscriptions, Black Friday planning and agency operations.
No publication dates are included with the posts, so their span cannot be established.
Two posts describe the same accordion case in near-identical wording and are counted as one result.
All revenue, test-count and lift figures are the author's own reported client results and are not independently verified here; most client brands are identified only by category and size.
The one measured audience figure is a LinkedIn following of 6,823, observed 9 October 2026.
The posts describe consulting and agency work rather than sponsored placements — no rates, deliverable lists or sponsorship terms appear in them.
Not stated in the posts: post dates, any earlier work, books or shows, engagement data and audience demographics.
