An investigation on Look Into starts with one question about a claim that shoppers rely on, such as whether a 20,000mAh power bank really delivers 20,000mAh. It ends with a dated answer, the evidence behind it, the working, and a plain statement of what we could not check. This page sets out the rules every investigation follows, so you can judge the findings for yourself.
What every investigation shows
- The question, and the short answer first.
- The sample: what we read, how we chose it, and how many.
- The dates on which every listing, specification and source was read.
- A label on every figure saying where it came from (see below).
- The working: every calculation shown in full, so it can be repeated.
- What we could not check, stated as plainly as the findings.
- Sources, linked, with the date each was read.
- A change log recording every correction or re-count after publication.
How we label evidence
Every figure in an investigation carries one of these labels:
- Maker's claim: what a seller or maker says in a listing or on a box, quoted as written.
- Maker's figure: a specification a maker publishes, such as a phone's battery capacity.
- Calculated: our arithmetic on figures that are labelled and sourced elsewhere on the page.
- Independent testing: results from a named laboratory or test body, linked.
- Look Into testing: our own measurements. None has been published yet. The first test's protocol is published ahead of any results: standby power of ten household devices.
How we choose a sample
When we audit retailer listings, we usually read the first 20 unsponsored results for a stated search on Amazon UK on a stated day, in the order they were returned, and open every product page. We leave out sponsored results, so the sample is what the search returns rather than what was paid to appear. We name the search and the date so anyone can repeat it, knowing the results will differ on another day.
Where it matters, we read each listing in two scopes: first the title and bullet points a shopper sees first, then everything the seller wrote. When the two disagree, we report both, because the gap is often the finding. Text inside product images is not counted, because a shopper cannot search it and a screen reader cannot read it; we say so on the page.
A sample like this describes those listings on that day. It is not a sample of the whole market, and we do not present it as one.
Claims about what is missing
Many of our findings are about what a listing leaves out, such as "0 of 20 state a rated capacity in text". A finding like that is tied to its date. A missing figure is a gap in the listing, not evidence that the product falls short. If we count again and the number changes, we add the new count with its date and keep the old one in the change log.
Naming companies
We name products and sellers, because a finding a reader cannot check is not much use. We describe what a listing or a maker says, at a date. Before we say a claim does not hold, we check the maker's own published figures or claims where they exist. We do not say anyone has broken a rule unless the evidence shows it, and we say when a gap is an omission rather than an error.
Right of reply
Any company named in an investigation can send us a response at hello@lookinto.co.uk. If it shows that a figure is wrong, we correct the page and log the correction with its date. If it disputes our reading of the evidence, we add a fair, dated summary of its response to the page.
No commercial links
Investigations, claims audits and cost indexes recommend no product, so they carry no affiliate links, no quote forms and no sponsored features, and earn Look Into nothing. Where an investigation points to one of our buying guides, it says that the guide does use affiliate links. How the rest of the site earns money is set out in how we make money.
Records and re-checks
We keep the record behind every count: the list of listings read, the figures as we found them and the calculations, so a finding can be checked or re-run. Standing indexes are re-checked on the dates shown on each page. When a re-check changes a figure, the new figure goes in with its date and the earlier one stays in the change log.
AI tools
We use AI tools, including Claude, to help read listings, run calculations and draft pages. The evidence rules on this page apply to everything they produce: every figure is labelled and sourced, every calculation is shown, and Jacob Whitmore reviews each investigation before it is published.
Corrections
If you think a finding is wrong, email hello@lookinto.co.uk with the page and the figure. We check it against the source and reply. Corrections are dated on the page; our full corrections policy is in the editorial policy.
Citing our findings
You are welcome to quote or reproduce our findings and figures with a link to the page they come from. Each investigation and index carries a suggested citation. The method behind our running-cost figures is on our methodology page.
Published 1 November 2026. Look Into is published by Whito Ltd, company number 10918465. Editorial decisions are made by Jacob Whitmore.