01
Homepage structure
Rebuilt to move a visitor through the store instead of leaving them to work it out.
- Hero section
- Trust bar
- Brand discovery
- Featured products
- New arrivals
Case study 01 — Shopify ecommerce
Table TennisUSA
A Shopify store selling table tennis equipment across the United States. In July it completed 234 orders. In the thirty days to 4 November 2026 it completed 615. Every figure on this page opens as the screenshot it came from.
Open the live store
The work is live — judge it there
01
Proof
Same store.Two and a half timesthe orders.
Completed checkouts, four consecutive periods. Counts rather than percentages, for a reason explained below.
Completed online-store checkouts, as recorded in the store’s Shopify Analytics. Online store only — this excludes the store’s other sales channels.
+163%
Completed orders, July against the thirty days to 4 November 2026
+33%
Carts over the same period — a third more carts, two and a half times the orders
26% → 52%
Share of carts that became an order
50% → 93%
Share of carts that reached checkout at all
Every period, every stage
| Period | Sessions | Carts | Checkouts | Orders | Conv. rate |
|---|---|---|---|---|---|
| Jul 1–31, 2026 | 16,349 | 899 | 466 | 234 | 1.43% |
| Aug 1–31, 2026 | 43,520 | 1,076 | 608 | 296 | 0.68% |
| Sep 1–30, 2026 | 18,037 | 1,084 | 719 | 353 | 1.95% |
| 5 Oct–4 Nov, 2026 | 15,516 | 1,194 | 1,113 | 615 | 3.96% |
Read the Orders column. It rises every period. The conversion-rate column does not behave the same way, and the two call-outs below explain why.
Why this page leads with orders, not conversion rate
1 — August’s traffic was not real
August recorded 43,520 sessions against July’s 16,349. Almost all of the increase came from one place: Singapore went from a handful of sessions to 21,500, and desktop sessions rose 268%. By September it had collapsed back to 789.
Conversion rate is orders divided by sessions. Roughly 25,000 sessions that were never going to buy anything pushed August down to 0.68%. The store did not get five times worse in August, and it has not become five times better since. Anyone who quotes 0.68% to 3.96% at you, including me, is quoting a number with a hole in it.
2 — Shopify changed how it counts sessions
A notice on the analytics page reads: “Beginning Sept 21, 2026, sessions and related metrics may shift as Shopify Analytics enables better visibility into real shopper activity.” A Definition change marker sits on the September chart at that date.
So any conversion rate spanning 21 September is not a like-for-like comparison with one before it. Order counts are unaffected, which is why they are the headline on this page.
What I will and will not claim.
What I claim: completed orders went 234, 296, 353, 615 across four consecutive periods, carts rose about a third over the same stretch, and the share of carts that reached checkout went from half to almost all. Every one of those is on a screenshot on this page.
What I do not: that every order is down to me. Traffic mix and seasonality move month to month, Black Friday trading sits inside the last window, and no single change isolates a cause.
And what this page deliberately leaves out: revenue. The store’s sales figures are the client’s to publish, not mine, so you will not find a dollar amount anywhere on this site.
Project overview
A big catalogue that was hard to shop.
In July 2026 the store pulled 16,349 sessions and converted 234 of them. Only 5.5% of those visitors ever reached a cart.
That is not a traffic problem and it is not a pricing problem. It was a business losing orders inside its own catalogue, and then losing a second batch between the cart and the card.
02
Diagnosis
Where the orderswere leaking.
July 2026, before the work. Of every hundred arrivals, five reached a cart, and half of the ones who reached checkout never finished.
July 2026, before the work
Sessions
16,349all traffic
Added to cart
8995.5% of sessions
Reached checkout
46652% of those carts
Completed checkout
23450% of checkouts
Each bar is that step’s conversion from the step immediately before it. Counts exactly as Shopify recorded them for 1–31 July 2026.
Two separate leaks, two different fixes.
Before the cart. Only 5.5% got that far. People could not find the right product fast enough, and nothing told them why to trust the store with a card.
After the cart. Half of everyone at checkout left. In equipment retail that is usually shipping cost arriving as a surprise at the final step.
The same funnel now
Sessions
15,516all traffic
Added to cart
1,1947.7% of sessions
Reached checkout
1,11393% of those carts
Completed checkout
61555% of checkouts
5 October to 4 November 2026. The first leak narrowed a little. The second one closed almost entirely: 93% of carts now reach checkout, against 52% in July.
03
The shipping fix
Shipping stopped beinga surprise at the end.
Two changes aimed squarely at the second leak.
Shipping options were rebuilt around what customers actually buy. A rubber sheet and a full-size table are not the same parcel and should never have been offered the same way.
Then the threshold was made visible. A free shipping progress bar sits in the side cart and at checkout, counting in real money toward the $75 mark.
It does two jobs. Removes the late surprise that was killing checkouts, and gives a shopper a reason to add one more item rather than leave. Both states below, from the live store.
04
The problem
Nine things standingbetween a visitorand an order.
Everything below was documented before any design work started. The list is what the project was scoped against.
01No conversion-focused homepage layout
02Weak product discovery for brand-loyal players
03Thin trust signals on the pages that needed them
04Unclear variant selection on equipment
05Shipping cost arriving as a surprise at checkout
06Broken breadcrumbs and SEO heading structure
The goal was to turn the store into a structured, high-trust shopping experience while keeping the Shopify setup scalable and maintainable — so the client could keep running it without me.
05
The solution
What actuallychanged.
Three areas of work. Each addresses specific items on the problem list rather than a general sense that things could look better.
02
Trust signals
Placed early, where hesitation happens, rather than buried in a policy page.
- Authentic global table tennis brands
- Professional quality equipment
- Secure checkout
- Trusted by players
03
Product page improvements
Reworked around the moment of decision: compare the specification, then commit.
- Clearer layout and spacing
- Improved product image presentation
- Trust messaging under the product price
- Improved quantity selector and purchase area
The homepage, in sequence
Five sections, in the order a visitor needs them.
01
Hero
What the store is, in one screen.
02
Trust bar
Four reasons to believe it, before any product.
03
Brand discovery
Butterfly, GEWO, Stiga, Victas — players shop by brand.
04
Featured products
Proof the catalogue is deep and current.
05
New arrivals
A reason for returning customers to look again.
Each section answers the question the previous one raises. A visitor who lands cold learns what the store is, why to trust it, which brands it carries, what it sells, and what is new — without having to search for any of it.
My role
Design and implementation, both.
I was not handing files to a developer. The design decisions and the code that shipped them were the same job.
- UX design improvements
- Shopify theme customisation
- Homepage restructuring
- Product page optimisation
- Conversion rate optimisation
- SEO structural fixes
- Custom feature implementation
What this means
The redesign did not just change how the site looks.
It changed how quickly people find a product, how much they trust the store, and whether shipping ambushes them at the last step. That is what a conversion rate is actually measuring.
And it stayed a Shopify store the client can run and extend themselves.
Next project — 02
GEWO USA
Contact
Want this run on your store?
Send the URL and your current conversion rate. I will tell you where I think the orders are leaking before you pay me anything.