Reviews and voice search: how ratings decide who gets recommended
- Reviews
- Voice search
Reviews help decide which business a voice assistant recommends, because an assistant gives only one answer and reputation is one of the few signals it can use to tell similar businesses apart. Assistants look at your average rating, the number of reviews, how recent they are and whether you respond as the owner — and sometimes they read your rating out loud to the customer. You do not need a perfect 5.0: a credible 4.6 with fresh reviews beats a 5.0 that has stood still for years.
Why do reviews weigh so heavily in voice search?
Ask Siri or Google Assistant where to eat well around here and you get no list — you get one answer. To pick that one answer, the assistant needs signals that tell businesses apart. Complete, consistent business details get you onto the shortlist — but in a town with three hair salons that all have their data in order, something has to settle it. That something is very often reputation.
And reviews do not stay behind the scenes. Assistants regularly read ratings out loud: "The Corner Barbers in Leeds, rated 4.6 stars." The customer hears one business and one score, with no comparison next to it. At that moment, your average is your entire shop window. AI assistants go a step further: they summarise what people write about you — "known for friendly service and honest prices" — and build their recommendation on that.
Which review signals do these systems look at?
No platform publishes its exact recipe, but four signals come up everywhere:
- Your rating. The average everyone knows. Important, but less decisive than most owners think.
- The volume. Ten reviews say something; a hundred say much more. The more voices, the more reliable the picture — for people and for algorithms.
- How recent they are. A steady stream of fresh reviews says: this business is alive today. A pile of glowing reviews from four years ago mostly says something about four years ago.
- Whether you respond. An owner who replies is visibly active and engaged. Future customers read that signal — and so do the systems judging your listing.
Notice what is not on the list: perfection.
You do not need a perfect 5.0
Quite the opposite. A 4.6 built on plenty of recent reviews is more credible than a spotless 5.0 from a handful of old ones — to customers, who tend to raise an eyebrow at a perfect score, and to systems that weigh volume and recency. One poor review is not a disaster. The pattern counts, not the outlier.
That is liberating news if you lose sleep over every star: you do not need to chase a flawless record. You need to make sure genuine, recent voices keep coming in.
How to ask for reviews the right way
Waiting for customers to write a review on their own is slow: it is mostly the unhappy ones who reach for the keyboard unprompted. So asking is fine — if you do it well:
- Ask at the moment of delight. The customer at a barbershop in Leeds smiling into the mirror, the regular at a café in Manchester coming over to say how good the lunch was — that is the moment. Not three weeks later by email.
- Make it effortless. A QR code by the till or a direct link sent by message makes all the difference. The fewer the steps, the more people actually follow through.
- Never buy or fake reviews. It is against every platform's rules, and it shows: a sudden wave of five-star reviews in the same week, all in the same style, from profiles with no history. Platforms filter them out, customers see through them, and you put your whole profile at risk.
- Spread it across the year. Asking for ten reviews a month beats a hundred in one campaign week. A steady rhythm is exactly the signal you want to send.
What do you do with a negative review?
Respond — calmly and concretely. Thank them for the feedback, name what went wrong without dancing around it, say what you are changing and offer to sort the rest out offline. No blame, no sarcasm, no public back-and-forth.
Why it pays off: future customers do not just read the complaint, they read how you handle it. A calm, concrete reply under an angry review says more about your business than ten five-star reviews. And to the systems judging your listing, every response is a sign of life: here is an owner who looks after their business.
Look beyond Google
Google is the most important review platform, but not the only one assistants look at. Apple Maps has its own ratings system — and if you claim your business through Apple Business Connect, you can respond there too. Sector platforms count as well: booking sites for restaurants, trade platforms for building and repairs, appointment sites for hairdressers and wellness.
For AI assistants this goes even further: ChatGPT and its peers read what is said about you across the whole web and melt it down into one verdict. So check regularly what ChatGPT says about your business — your reputation on a platform you never look at still counts there.
Building your entire reputation on one platform also leaves you exposed. One policy change or one wrongful takedown and you are left with nothing. Spreading out is the safer route — and it makes your story more credible to every system, because here too the rule holds: sources that confirm each other build trust.
Curious how your reputation signals look today, next to your business details? The free RankPrima Scan shows you in a few minutes where you stand strong and where the first work lies.