In an era where consumers trust each other far more than they trust brands, the smartest marketers are learning to hand over the mic. Joe Johnston, Head of Experimentation at Launch, takes readers inside the new frontier of acquisition; one fuelled not by guesswork or polished brand narratives, but by the raw, unfiltered enthusiasm of real customers.

Being talked about by customers, in a positive light, is one of the most powerful acquisition strategies for a brand. 

The strength of social proof as a psychological bias - through great reviews, user generated content, or mentions in the media – makes a brand more trustworthy. As well as removing the biggest conversion killer of all - doubt.

We all know as marketers that our role is to connect a brand with its target audience on a rational level. And perhaps more importantly, emotionally too (especially as the price point increases). 

Which is why it is becoming increasingly vital to understand how to turn a routine interaction into an experience someone is compelled to rave about. As well as  recognising the brand experiences that will drive a customer to take five minutes to write a glowing report on Trustpilot or upload their take on TikTok.

It’s simpler than a brand may think

What most have yet to realise is that they are sitting on a goldmine of brand-customer connection stories. Specifically those who have over 1,000 reviews (for statistically meaningful patterns volume is key, otherwise the sample size is too small).

For those who do, every four- or five-star customer review that a brand collects includes a real-life reason to buy. A story that explains in clear terms the exact brand experience that led this person to leave a positive rating. 

This is gold-dust for many reasons. For one, it’s an actual lived experience, rather than a hypothetical  study. Often during consumer research, participants are asked questions about what their intent is to buy and why. ‘When do you next plan to go on holiday to x?’ ‘How much would you pay for y?’ This is good directional data, but there’s often a big gap between what consumers say they will do, and what they actually do. 

Reviews, on the other hand, are from verified customers from a real life event. They talk passionately about the parts of the brand experience which are most important to them, without being biased by the researcher’s questions or framing. Not to mention the fact that they are often inclined to compare a brand to an alternative (potentially a key competitor).

On top of that, there’s also normally thousands of reviews, a scale which reduces the risk of one or two outliers skewing results. Best of all, it captures language that customers naturally use. And those localised or multi-segment brands gain extra value from regional nuances. Qualitative data doesn’t get much better than that. 

The problem is, up until recently, the job of categorising, parsing and analysing thousands, or sometimes tens of thousands of individual reviews would take a single researcher months of mind-numbing labour. 

For years, the qualitative value of customer reviews was locked away. Even the most well-funded insights team couldn’t justify months of manual coding. As a result, many brands focused on star ratings and ignored the stories beneath. Now, all of a sudden that has changed.

Which is why a good insights team with expertise is needed to extract the data safely. Over the past twelve months, at Launch, we’ve conducted half a dozen customer review mining projects for brands, some with up to 80,000 reviews. 

AI is also playing a critical role here

It's helping insights teams such as ours to unlock the potential qualitative value stored in customer reviews. 

With the right prompts, an LLM will uncover the most discussed themes and topics, as well as overall sentiment distribution. This enables the identification of patterns and trends in customer feedback over time and across demographics. All in the space of minutes. It's become vital to reshaping an acquisition messaging and testing strategy.

The output of these projects is a messaging framework which uses authentic phrasing, mirroring the customer’s natural language and concerns. Often in reviews, customers ‘code switch’ using phrases brand experts could never think of. 

However, just to caveat. It's important to take into account a couple of rules when using AI for review mining. Firstly, never use personally identifiable information and keep customer data safe with strong privacy protections and enterprise-grade security. Additionally do not use conversations for model training. 

Data should not end up appearing in another users’ chat conversation. Which is why it is also vital to make sure that data is cleaned and structured to enable chosen LLM/s to understand it. 

In terms of prompts, it’s essential for a brand to experiment with trial and error. Never take the first answer at face value. Check the source data to make sure it’s not hallucinating. And nail down the prompts which will repeatedly give you a trustworthy response.

This approach works because we’re wired this way

Humans subconsciously like and trust those who mirror their speech patterns. Ads that mirror audience vocabulary create rapport and trust. Brands that “sound like me” feel more relatable and fluent, reducing psychological distance. So consumers are more likely to buy.

The best part of customer review mining is that there are no upfront research costs.  It’s simply about a brand taking advantage of all the hard-earned positive brand experiences that it has been facilitating and meticulously collecting for years.  Surely, there’s no better way to attract and acquire new customers, who are more likely to stick around, than to listen carefully to a brand’s happiest customers and let them guide more effective and impactful advertising? 

Put simply, mining customer reviews is creating a brand experience that will resonate on such an authentic level, that customers will want to talk about it. 

So, what are you waiting for?