Google Reviews and Your Website Are One System

Open the Google Business Profile of a well-run suburban electrician and you'll often find a rating in the high fours built off a couple of hundred reviews, each one written by someone who stopped what they were doing to type it. Then open the same business's website and there's nothing — or a "Testimonials" tab holding three unattributed quotes that have sat there since launch.

That gap is the problem. Reviews and the website get run as two separate projects, and the most persuasive asset the business owns ends up parked on a platform it doesn't control. Our position: they're one system. Google is where the proof gets collected. The website is where it has to work, because that's where somebody sits with your phone number in front of them deciding whether to use it.

Joining them up is cheap. The expensive part is how businesses usually get there — the small incentive that breaches Google's policy, the widget that adds a second to the mobile load, the aggregate rating markup that says nothing.

The two jobs a review has to do

On Google, the rating and the count are discovery inputs. They sit in the local pack next to your competitors, and volume, recency and content all feed into whether you appear at all. For that job a two-word "great service, thanks" is genuinely useful.

On your website the job is different. The visitor is already there and has decided you're plausible. What they want now is evidence you've handled a situation resembling theirs, and the two-word review does nothing for that. The one that earns its place names the job, the suburb, the thing that went wrong and how you dealt with it — a case study written by someone with no reason to flatter you.

Some visitors will open a tab and read your profile. Plenty won't. For that group, the reviews you didn't put on the page may as well not exist. We've covered what makes social proof land on an Australian website before: unattributed praise reads as marketing, a named review with a suburb attached reads as a fact.

An ask system that doesn't breach anything

Most businesses with thin review counts don't have a reputation problem, they have an asking problem. Nobody asks, or somebody asks three weeks late by email from an address the customer doesn't recognise.

Ask at the moment of relief — when the customer can see the result and feels good about it. For a trade that's standing in the finished room, not when the invoice lands. For a clinic or a workshop it's at the counter on the way out. Ask then, in person where you can, and follow with a text containing the direct review link. The text lands on the device that will write the review; email asks someone to change context, and most people don't.

What you cannot offer

Google's Business Profile policies prohibit incentivised reviews outright — a discount, a free coffee, a raffle entry. Reviews collected that way are subject to removal, and repeat behaviour can cost you the profile.

The larger risk sits under Australian consumer law. Fake, incentivised or manipulated reviews are misleading conduct, and the ACCC has named reviews an enforcement focus repeatedly. Since November 2022 the maximum penalty for a corporation breaching the consumer law has been the greater of $50 million, three times the benefit obtained, or 30 per cent of adjusted turnover across the breach period.

The instructive Australian case isn't about fake reviews. Accommodation operator Meriton was found to have masked guest email addresses so TripAdvisor review requests never reached guests likely to complain. No invented reviews, no paid ones — just the real negative ones quietly prevented from happening. The Federal Court treated it as misleading conduct and imposed a multi-million-dollar penalty.

Which makes review gating the trap

Review gating is the tidy system where you survey customers first, send the happy ones to Google and route the unhappy ones to a private feedback form. Plenty of review software still sells it as a feature. It breaches Google's policy, and structurally it's the conduct that put Meriton in front of the Federal Court: engineering the public record so it shows the good half. Ask everyone or ask nobody, same request every time, including the customers you suspect will be lukewarm.

Recency does more work than the total

Google displays relative dates, which turns your profile into a liveness check. A 4.9 from 180 reviews where the newest is fourteen months old reads as a business that has wound down, and the visitor can't tell quiet from closed. A steady trickle also changes what a bad review costs you: one angry review on top of eleven from the last quarter is context, while the same review on top of silence is the first thing every prospect reads for a year. Fifteen arriving in a single afternoon, meanwhile, looks like exactly what it is.

Replying in public is conversion copy

The reply to a bad review is one of the most valuable things your business will publish, and almost everyone writes it to the wrong audience.

The reply isn't for the person who left the review. It's for the person reading it six months later, deciding whether to call you.

That reframing sets the structure. Acknowledge the experience without grovelling. Include one specific, non-identifying detail proving you know which job this was — evidence of a real operator, not a template. Say what changed. Then offer a path offline, with a name and a number.

What never works: relitigating the facts, "we have no record of you as a customer", anything with legal temperature in it, or a defensive wall longer than the complaint. The reader isn't adjudicating who was right. They're watching how you behave when something goes wrong, because they're imagining it going wrong for them. Reply to the good ones too, in two lines — keyword-stuffed replies are obvious and they cheapen the profile.

Embedding reviews without lying in your markup

The common approach is a third-party widget: paste a script tag, get a carousel. It works, and it costs you — its own JavaScript, fonts and stylesheet, often several hundred kilobytes, frequently rendering last so the layout jumps as it lands. On a mid-range Android over mobile data, a review carousel is a strange thing to spend your page budget on.

The better build fetches reviews server-side, caches them, and ships plain HTML with the page. No third-party script, no layout shift, and the reviews are still real and current. One constraint: Google's Places API returns a maximum of five reviews per location, which is why nearly every widget rotates the same five. Beyond that you're into Business Profile API access, or curating manually — quoting accurately, attributing properly, linking to the profile so anyone can verify. Manual curation is what we'd usually recommend, because you choose the reviews rather than whichever five Google surfaces.

Then the markup, where a persistent myth costs people time. Since 2019 Google has not shown review rich results for self-serving reviews — reviews about your own business, on your own site, marked up on your own LocalBusiness or Organization schema. Google's review snippet documentation is explicit. The stars beside a local business come from its Business Profile, not from anything in your HTML.

So a hand-written AggregateRating on your homepage is doubly pointless. No rich result, and if the number doesn't match a real countable set of reviews, you've published a false claim about your own business in machine-readable form — a consumer law problem with a schema wrapper on it. The SEO fundamentals for a small business site cover what is worth marking up instead.

If you're unsure what your current widget costs the page, run the URL through our free audit — it reports what loads, in what order, and how long the mobile render takes.

Where reviews belong on the page

Placement is where most of the conversion difference lives.

A dedicated reviews page is worth building, last, for the people who want to dig — just never treat it as where reviews live. For trades this placement work is usually the cheapest conversion improvement available, which is why it's baked into how we approach trades and home services sites.

Health services play by different rules

If you run a physiotherapy practice, a psychology clinic, an optometrist or a cosmetic clinic, most of the above doesn't apply, and getting it wrong is a regulatory matter.

The National Law prohibits testimonials in advertising a regulated health service, and AHPRA treats a review referring to clinical care as a testimonial. You aren't responsible for unsolicited reviews on platforms you don't control — but the moment you embed them on your own site, or ask patients to write them, you're using them in advertising. Comments that don't touch clinical care, such as parking or reception, sit outside the definition, though the safe reading is to stay clear of that boundary rather than test it.

So a regulated health site builds trust from other material: practitioner credentials and registration details, real clinician bios, clear fees, honest waiting times, photography of the actual rooms. Harder, and a differentiator — that clinic reads as considerably more professional than one leaning on a star carousel it shouldn't have installed.

Veterinary practices are the common point of confusion. Vets are regulated by state boards rather than AHPRA, and client reviews are fine.

FAQ

Can I offer a discount for a Google review?

No. Google's Business Profile policies prohibit incentivised reviews of any kind — discounts, free products, prize draws — and reviews collected that way can be removed. Separately, an undisclosed incentivised review can be misleading conduct under Australian consumer law.

How many Google reviews does my business need?

There's no threshold that switches something on. What matters more is whether reviews are still arriving — 40 reviews including several from the last two months usually reads better than 200 that stopped eighteen months ago.

Can I delete a bad Google review?

Not directly. A business can't remove reviews from its own profile. You can report one that breaches Google's content policies — spam, conflicts of interest, off-topic content, harassment — and Google may remove it. For a genuine unhappy customer, your only lever is a good public reply, which is worth more anyway.

Do reviews on my website help my SEO?

Indirectly. Review content adds relevant language to the page, but marking up your own reviews with AggregateRating will not produce stars in Google's results — self-serving review snippets have been ineligible since 2019.

Twenty minutes a week

None of this needs a project plan. It needs a repeating slot in the calendar.

Once a week, open the profile and read anything new. Reply to everything inside 48 hours, writing each reply for the next reader. Then check the asking habit happened — if no reviews arrived and you completed eleven jobs, the ask isn't being made, and that's the only thing to fix this week.

Once a month, pull the best two or three new reviews into the site and match them to the pages they belong on: the service they describe, the suburb they name, the form they should sit beside. On a well-built site that's a ten-minute content change, not a development ticket.

Once a quarter, check the plumbing. Is the review link on the invoice still working? Is the profile even appearing when someone searches your category locally — because if it isn't, the reviews aren't reaching anyone, and that's a Business Profile visibility problem rather than a reputation one.

The businesses that win at this aren't the ones with the highest ratings. They're the ones whose proof sits in front of the visitor at the moment of decision instead of two clicks away on someone else's platform. For a straight read on where reviews belong on your site, book 20 minutes or call us on 0421 933 907 — bring your profile URL.