Friday, 20 November 2026

In 2026, scoring 4.9 on Google is simply not enough

For years now, savvy businesses have done just about everything they can to achieve a great Google score; in 2020, that was anything around 4.5; by 2023, that had elevated to 4.8, and this year it is definitely 4.9. How so? If you conduct a search for any professional or service business - legal, financial, estate agency and so on - you will find the top-ranked business now scores 4.9.

Here's an example that is repeated almost everywhere:




Another interesting point: the search string, especially where people are using AI search, is lengthening for the first time in many years. Back in the day, potential customers were forced to use pretty long search strings to pinpoint the business best suited to their needs. For example, 'estate agency in Kingsbury North London'. With advances in search technology, this string was reduced right down to 'estate agent' or 'estate agent near me'. None of these search strings was qualitative, simply because Google could not return 'best'. Until now. 

So how does Google's Gemini AI (or ChatGPT or Grok, for that matter) establish its qualitative results? As with any AI, it searches for evidence. What evidence? We will let Gemini answer that...




The first two - 'Relevance' and 'Distance ' - are fixed. So nothing will differentiate the dozen or so estate agents in the local high street; they are all estate agents ('Relevance') and all in the same geographical area ('Distance'). Now we begin to see the importance of the third and fourth resources AI uses. And if we need evidence that AI is referencing reviews, both Google and those hosted on the business's own website*, then look no further than the AI deep dive results:



You might even notice that Gemini's first paragraph above directly quotes this individual review (shown right under the above paragraph in the Gemini result):



We blush! But seriously, Google's Gemini faces stiff competition from other AI platforms; it would be commercial madness not to mine down into the business's own reviews as well as its Google reviews when formulating search results. Simple duplication - hosting the same reviews displayed on Google on the business's own website - besides being commercially questionable (data being cash on the balance sheet for most businesses) - is denying Google and other AI a resource that can be the difference between coming top and being shown as an also-ran.

Here is ChatGPT:




But back to our original point: 4.9 not being enough. We have dealt with the fact that AI searches will always highlight the businesses with the top scores from all sources for any given search, and regular readers (and HelpHound members) will know that moderation contributes to that in more ways than one (giving the business the confidence to go out and invite reviews in the knowledge that those that are factually inaccurate, potentially misleading or just plain unfair are highly unlikely to see the light of day as a final published review, either on the business's website or on Google), and this is where our next point comes from: a business can score well, but still, now in the days when online reviews are reaching maturity (in pure volume, if not always in content!), still harbour unfairly offputting reviews, and these, when viewed and read by a potential customer, can do considerable damage.

Let's take Octopus Energy, a business many of us have first-hand experience of, as an example. It has reviews on Google:




Where nearly 600 reviewers rate it at 1*

And it proactively invites reviews on Trustpilot:






Where just over 27,000 rate it one star. 

That would be fine if individual consumers and AI couldn't mine the individual reviews. But they can, and do. They also ask AI search questions like this:




We know for a fact that ChatGPT is referencing more sources than simply Google or Trustpilot reviews, but when one reads the individual reviews, as many cautious consumers will, then the comments in bold begin to ring true.

So, as you can see, a great review score can still mean an awful lot of underlying negative comments, much of it inaccurate or unfair, that could have been resolved with a moderated system*. Resulting in an even better score and far fewer negative returns when the reviews are read, whether by AI or an individual.


*Moderation also improves the quality of reviews; when the reviewer understands the value the business places on the content of the review, not just the score, the average word count of the reviews invariably increases. This results in very convincing individual reviews. It is interesting to see businesses like Octopus taking the opposite tack, where the objective is to achieve a high score at all costs; so many of their reviews are not of their core service, but of a brief interaction with a member of staff. This is to confuse review management and consumer feedback. This, to our mind, degrades not just the impression given by a great overall score, but the entire value of such a score.


Further reading



No comments:

Post a Comment

HelpHound is all about feedback, so please feel free to comment here...