Skip to content

How we test and review

A company that writes guides and comparisons has a stake in how they come out. So do we. These are our rules.

How does Alfred keep comparisons fair?

Alfred

Every comparison says that we sell Local SEO, Reputation management, Reception, Growth and Multi-location, shows its method and test dates, and sets out what each provider does well. Alfred is never ranked first by default. Every number has a source: our own research with its sample and date, or a linked primary source. AI answers show why: US ChatGPT's review counts matched Google's in 22.9% of checks in October 2026 (method).

We sell what we write about, and say so

Alfred sells Local SEO, Reputation management, Reception, Growth (web development, SEO, advertising and social media) and Multi-location. Every comparison, and every guide that covers a category we sell in, says so at the top. You should know who is writing before you read what they wrote.

The same goes for numbers. AI assistants pass figures on without checking them. When ChatGPT showed review counts for US contractors in October 2026, they matched Google’s own count in 8 of 35 checks, and most were far lower, traced to Yelp’s figures and to cards attached to the wrong listing (method). A page that an assistant may quote has to carry numbers that survive a check. Every number on our pages is traced to its source before it goes up.

The rules for every number

A number on our pages comes from one of three places, and from nowhere else:

  • Our own research, with its sample, its fieldwork date and a linked method page that sets out the sample, the tools and the limits.
  • A linked primary source: legislation, a regulator, the platform’s own documentation, with a date.
  • A worked example, clearly labelled as made-up numbers.

References to studies without a name and a link do not appear. That applies to what we say about our own services too: what we say Alfred does is what Alfred does today, for paying customers.

How we use AI, and where people come in

Drafts are written with AI assistance from three inputs: our approved fact sheet on what Alfred does, a brief with the questions the page has to answer, and the data pulled for that page. Nothing outside those three may appear in a draft. Where a fact is missing, the gap is marked, and a person fills it with a source or cuts it.

Then people read. An editor traces every number to its source, in titles, descriptions and captions as well, as Google’s guidance on AI-generated content asks (Google Search Central).

We do not publish pages in bulk to show up in search or AI answers. Google’s spam policies count content produced at scale to influence rankings or AI answers as a breach (Google Search Central), and a page that adds nothing to what already exists does not get built.

What we write about ourselves

  • Product claims describe features in use today.
  • The price is set per location by what you need and given in writing after a demo, before you decide (pricing: UK, US, UAE).
  • No client names, reviews of Alfred or results appear without the client’s written approval, and the businesses in our page designs are labelled as examples.
  • No promises of rankings. Google and the AI assistants decide, and we show patterns.
  • We ask every customer for a review, never only the happy ones, and we never write or buy one. Fake reviews are banned in the UK under the Digital Markets, Competition and Consumers Act 2024, with fines of up to 10% of global turnover (CMA). In the US the FTC’s rule on consumer reviews and testimonials has banned them since 21 October 2024 (Federal Register).

Corrections and updates

If you find an error, write to hej@getalfred.se with the page address and what is wrong. Factual errors are corrected, and the page gets a line saying what changed and when. When a figure in our research changes, every page that uses it is updated the same week.

See where there is more to win.

Type your business name and get a free check in about a minute.

No credit card · No meeting needed