How we test
Morningfold's value depends on the fact that the picks are real. This page documents how we get there, what we test, how we test, and where the limits are.
The basic rule
No product is featured before it has been used.
Where the product is software, it has been installed and used as a primary tool, not skimmed during a press demo, not summarised from a marketing site. Where the product is hardware, it has been bought (or borrowed and returned) and lived with, not merely unboxed.
Where the product is a service, banking, broadband, mobile, energy, we have opened a real account in James's name, moved real money or used real bandwidth, and either kept the account or actively closed it. Both outcomes inform the verdict.
Testing periods by category
Different products need different windows to reveal themselves. Roughly:
| Category | Minimum testing window |
|---|---|
| AI tools (productivity, writing, research) | 2 weeks of daily use |
| Health and wellness apps | 4 weeks |
| Wearables and devices | 6–8 weeks (one full feature cycle plus return-policy window) |
| Banking and fintech | 6 weeks (covers a billing cycle and at least one customer-service interaction) |
| Broadband and mobile | 4 weeks (covers a full bill plus speed measurements at multiple times of day) |
| Mid-cycle upgrades to previously-tested tools | 1 week (sanity check only) |
If a tool moves to subscription pricing, changes its core model, or is acquired during the window, we restart testing.
Where the testers are
All primary testing happens from addresses on accounts. London (Greater London) and one rural location in the South West are the two main test bases. This matters for things like broadband (urban fibre vs FTTC), mobile coverage, delivery times, and warranty handling, all of which differ regionally.
When we draw on testers outside the UK (rare; usually for tools that aren't UK-released yet), this is disclosed in the article.
What we measure
The metrics depend on the category. A few constants:
- Real-money cost over a year, including taxes, not the headline price
- What happens if you stop using it (subscription lock-in, data export, etc.)
- Customer service response time, we deliberately log a real query and time the response
- What it does poorly, every product gets a "what's not good" section, and if we struggle to write one, the product probably needs more testing
What we deliberately do not do
- Press loaner reviews of devices without buying them ourselves where the price exceeds £200. Loaned devices are routinely cherry-picked. We trust our own purchase or nothing.
- Affiliate-driven category lists. When you see "the best [category] tools" articles online, the order is almost always set by commission rate. Our top picks are decided before any affiliate URL is added; we publish the verdict regardless of whether the winning product has an affiliate programme.
- Sponsored editorial. Brand collaborations are clearly marked as sponsorship in the article and the email, and they sit in their own template. No sponsored content rewrites editorial picks or appears in our "best of" coverage.
- Re-publishing of press release content. We don't republish vendor announcements as news.
What we will revise
If a product changes meaningfully, a price hike, a feature removal, a customer-service collapse, we update the article and the verdict, then send a brief note to subscribers in the next morning email. The original article carries a dated note explaining what changed.
If the verdict changes after publication, the article header is updated and the change is logged in a public footer. We don't quietly edit history.
Conflict-of-interest disclosure
- Morningfold's editor (James Walker) holds no equity in any AI, health-tech, fintech, or telecommunications brand featured. If that ever changes, the relationship is disclosed at the top of any relevant article and we do not publish on that brand for the duration.
- Morningfold has affiliate partnerships with several brands across all five topics. The complete list is on our Advertise & sponsorship page so readers can see exactly where commercial relationships exist.
- Morningfold accepts no payment for editorial placement or favourable ranking, ever.
Reader corrections
If you've used a product we covered and disagree with the verdict, please email hello@morningfold.co.uk. Genuine disagreements that change the verdict are logged in updated articles, with credit (anonymised if you prefer).
Last updated: 26 April 2026.