For two decades, the answer to "how do we rank" involved backlinks. They were the closest thing search had to a vote of confidence, and an entire industry grew around acquiring them.
Answer engines changed the weighting. Links still matter, because they still help a model find and trust a page. But when an assistant is deciding which company to name, review consensus does more work than link volume, and most companies have their investment the wrong way round.
Why the shift happened
A link is a signal that someone thought a page was worth pointing at. A review is a statement about whether a business is actually good, written by someone who used it. When a model is composing a recommendation it must stand behind, the second kind of evidence is more directly useful.
Reviews also have three properties engines love. They are numerous, so patterns are statistically meaningful. They are recent, so they say something about the business now. And they are specific, so they support precise claims: not just that a company is good, but that it is good at a particular thing, for a particular kind of customer.
What engines actually extract from reviews
Not just the star rating. Modern models read the text and pull out themes.
- Recurring specifics. If forty reviews mention that you turn work around quickly, the engine can confidently recommend you for urgent jobs.
- Who you suit. Reviews reveal customer type in a way marketing copy never does.
- Consistency over time. A steady flow beats a burst, which reads as either a campaign or a manipulation.
- How you handle problems. Thoughtful responses to negative reviews are strong evidence of reliability. A wall of perfect scores with no dissent reads as less credible than a strong average with a few honest complaints handled well.
The practical rebalance
If you are spending on link acquisition and doing nothing systematic about reviews, you are optimising for the previous era. A reasonable split for most businesses is to put the majority of the effort into a consistent review habit and the remainder into earning genuine mentions in sources your industry actually trusts.
By systematic, we mean a defined moment in your delivery process where the ask happens, every time, to every satisfied customer, on the one platform that matters in your category. Not a campaign. A habit.
A link says someone noticed you. A review says someone chose you and it worked out.
The line you do not cross
Never buy reviews, never incentivise them, and never write them. Beyond the legal exposure, which in Australia is real and enforced, engines are increasingly good at detecting the patterns that fake reviews produce: clustered timing, similar phrasing, reviewer accounts with no other history. Getting flagged does not just cost you the fake reviews, it devalues the genuine ones alongside them.
The slow, honest version compounds. The fast version eventually costs you the asset you were trying to build.