Strategy
How London fashion stores should test mobile checkout against Manchester
London vs Manchester mobile conversion fashion: compare device mix, baskets and returns to set sharper A/B testing priorities for British fashion retailers.
What to take away
- London vs Manchester mobile conversion fashion differs because London traffic skews newer iPhones on fast networks while Manchester holds a higher Android and older-device share.
- Basket sizes run larger in London, but returns rates are also higher there, so test revenue net of returns, not gross order value.
- One test can read both cities if you segment by device and region rather than running separate campaigns.
- Use ONS regional data for device and basket comparisons instead of assuming your analytics sample is representative.
Device mix differences between London and Manchester fashion shoppers
Device mix is the first thing to split when you compare London and Manchester fashion shoppers. London has the UK's highest concentration of newer handsets and fast mobile networks. Manchester, in North West England, holds a larger share of mid-range Android devices and older iPhones.
No public dataset splits device mix, basket size and returns rate by London and Manchester for fashion checkout. Pull them from your own analytics. In Google Analytics 4, build an exploration with region and device category, then add average purchase revenue and return rate. In Shopify, combine the sales by region report with returns data.
For illustration only, a typical fashion store might see iOS share around 60% in London and 45% in Manchester, average basket around £85 in London and £70 in Manchester, and returns around 25% in London and 18% in Manchester. Your own numbers will differ.
London vs Manchester device mix
London
- Dominant device
- Newer iPhone
- Network speed
- Fast
- Device age
- Newest handsets
- Viewport size
- Larger
Manchester
- Dominant device
- Mid-range Android
- Network speed
- Slower, mixed
- Device age
- Older iPhones
- Viewport size
- Smaller
That gap matters for mobile checkout testing. A payment step that renders cleanly on a recent iPhone can still break on a three-year-old Android with a smaller viewport. If your London team tests only on their own phones, they will miss the failures Manchester shoppers hit every day.
ONS population and community data gives the demographic backdrop: age profiles, household composition and regional spread differ enough between Greater London and North West England to shift device ownership. Use that data for regional consumer demographics rather than guessing from your own traffic. People, population and community - Office for National Statistics
Economic differences reinforce the split. Regional accounts show London and Manchester sit in different income and output bands, which feeds into what shoppers will pay for a basket and which devices they carry. Regional accounts - Office for National Statistics
For a London fashion store, the practical move is to pull device mix by region from your analytics, then compare it with ONS regional figures. If your Manchester traffic is under-represented in your test sample, your results will flatter London and mislead the buying team.
Basket sizes and returns behaviour in each city
London baskets tend to be larger. ONS regional gross disposable household income data shows London households have higher disposable income per head than North West England. ONS family spending data covers regional clothing and footwear expenditure. A denser store estate for click-and-collect also pushes average order value up.
Manchester baskets are smaller on average but more considered, with a higher share of repeat purchases from loyal local customers.
Baskets and returns by city
London
- Average basket
- Larger
- Returns rate
- Higher
- Return reason
- Multiple sizes ordered
- Key metric
- Net revenue per session
Manchester
- Average basket
- Smaller, repeat
- Returns rate
- Lower
- Return reason
- In-store returns
- Key metric
- Net revenue per session
Returns behaviour runs the other way. London shoppers return more, partly because they order multiple sizes and colours to try at home. Manchester returns are lower, but the reasons differ: fewer wardrobes, less occasion buying, and a stronger preference for in-store returns that staff can resolve without a courier.
This changes what you should measure. Gross conversion rate is a vanity metric if returns eat the margin. Track net revenue per session, return rate by region and return reason codes. A London test that lifts orders but doubles returns is not a win.
ONS retail industry data gives the benchmark for online share of retail sales, so you can see whether your own online mix is normal for the period. Retail industry - Office for National Statistics
Basket sizes and returns behaviour
London pattern
- Dominant device
- Newer iPhone, fast network
- Average basket
- Larger
- Returns rate
- Higher
- Return channel
- Courier and drop-off
- Test priority
- Checkout speed and payment
Manchester pattern
- Dominant device
- Mid-range Android, mixed network
- Average basket
- Smaller, more repeat
- Returns rate
- Lower
- Return channel
- In-store preference
- Test priority
- Delivery clarity and sizing
Setting A/B testing priorities for a London fashion store
Start from the decision event, not the tool. Define what a successful checkout looks like for a London shopper, then for a Manchester shopper, and only then pick the test. This is the discipline behind good audience research. audience research
A/B testing priority order
- Fix broken mobile checkout on London traffic
- Test payment options for older Android
- Test sizing and returns messaging for Manchester
- Test delivery speed claims against couriers
- Test guest checkout against account creation
Rank your priorities in this order:
Setting A/B testing priorities
- Fix broken mobile checkout steps on London traffic, because that is where the volume and the revenue sit.
- Test payment options that older Android devices handle without a redirect.
- Test sizing and returns messaging for Manchester, where returns are lower but sizing doubt still blocks checkout.
- Test delivery speed claims against real courier performance, not marketing copy.
- Test guest checkout against account creation, segmented by city.
Each test needs a sample size that survives a city split. Do not run a city test with a sample that only proves noise.
Choose your tooling with UK pricing and GDPR context in mind. UK teams often use Optimizely, AB Tasty or VWO for A/B testing, and Google Analytics 4 for segmentation.
A/B testing tools that store data outside the UK need a lawful transfer mechanism under UK GDPR and the Data Protection Act 2018. Use the UK International Data Transfer Agreement (IDTA) or the UK Addendum to the EU Standard Contractual Clauses.
Mobile checkout steps that fail on London traffic
London mobile traffic is fast and impatient. The steps that fail are predictable.
Mobile checkout failures to fix
- Address lookup loses form state
- Bank app redirect never returns
- Postcode validation rejects valid London codes
- Delivery slot pickers load slowly
- Promo code fields punish no-code shoppers
Mobile checkout steps that fail
- Address lookup that opens a third-party overlay and loses form state on return.
- Payment redirect to a bank app that fails to return the shopper to confirmation.
- Postcode validation that rejects valid London postcodes such as E1W or SE1.
- Delivery slot pickers that time out on congested mobile networks.
- Promo code fields that block checkout when no code is available.
For example, a redirect to a bank app such as Barclays or Monzo can drop the return URL on some Android WebViews. A London shopper on a newer iPhone may never see the issue.
Google's guidance on optimising websites for mobile is clear that speed, tap targets and form design drive mobile conversion. Optimise your website for mobile
Our mobile checkout optimisation checklist covers the repair stage in detail.
What Manchester data can and cannot tell a London retailer
Manchester data can tell you how a mid-range Android shopper behaves when the network is slower and the basket is smaller. That is useful stress-testing for your checkout.
It cannot tell you the London price sensitivity, the London returns rate or the London delivery expectations. Those differ enough that copying a Manchester winner into London is a gamble.
Use Manchester as a control region, not a proxy. If a change wins in both cities, it is likely a real improvement. If it wins only in Manchester, you have learned something about device performance, not about London shoppers.
Be careful with small samples. A Manchester test with a few hundred sessions will not detect a two per cent lift. State your minimum detectable effect before you start, and stop when you hit it.
Designing one test that reads both cities
One test can read both cities if you design it as a segmented experiment rather than two campaigns.
Segmented test launch checklist
- Primary metric is net revenue per session
- Return rate tracked by region
- Device mix reported per city
- Sample size set for smaller city
- No promo or seasonality overlap
- UK GDPR and DPA 2018 compliance
- Reviewed by someone who did not build it
Define your target audience as shoppers who reach the checkout decision event, then segment the results by region and device. That gives you a single test with two clear cuts.
Use this checklist before you launch:
- Primary metric is net revenue per session, not orders.
- Return rate is tracked by region for the test period.
- Device mix is reported for London and Manchester separately.
- Sample size is calculated for the smaller city, not the larger.
- No promo or seasonality overlaps the test window.
- Data processing complies with UK GDPR and the Data Protection Act 2018.
- Results are reviewed by someone who did not build the test.
For a Manchester baseline conversion of 3% and a target lift of 20% relative (to 3.6%), you need about 14,000 sessions per variant at 80% power and 95% confidence. If Manchester sends 7,000 checkout sessions a month, run the test for two months or accept a larger minimum detectable effect.
Run for a full week minimum, ideally two, to cover weekday and weekend behaviour. London and Manchester shop at different times, and a three-day test will favour one city over the other.
Then decide. If the change wins in London and holds in Manchester, ship it. If it wins only in London, ship it for London and keep testing for Manchester. If it wins only in Manchester, investigate device performance before you touch the London experience.
