{"id":110135,"date":"2026-09-30T09:00:00","date_gmt":"2026-09-30T16:00:00","guid":{"rendered":"https:\/\/www.crazyegg.com\/blog\/?p=110135"},"modified":"2026-09-23T12:06:06","modified_gmt":"2026-09-23T19:06:06","slug":"review-generation-strategy","status":"publish","type":"post","link":"https:\/\/www.crazyegg.com\/blog\/review-generation-strategy\/","title":{"rendered":"Review Generation Strategy: Why consistent positioning matters more than volume and star ratings for AI-assisted buyer journeys"},"content":{"rendered":"\n

Run an NPS survey, reach out to the happiest users, and offer them an incentive for a glowing recommendation. For years, that formula has been the standard for review generation in SaaS.<\/p>\n\n\n\n

However, traffic to sites like G2 and Capterra has declined significantly. Instead of visiting directly, buyers are using AI tools as an intermediary to consume reviews. They\u2019re asking specific questions, checking platform shortcomings, and triangulating with other sources to verify claims. <\/p>\n\n\n\n

What does this mean? In short, chasing lots of positive reviews from happy customers doesn\u2019t cut it any more.<\/p>\n\n\n\n

An overhaul of the old solicitation model is needed. Vendors that build detailed and honest bodies of customer evidence will be much more effective at shaping how LLMs understand and respond to questions about their products. <\/p>\n\n\n\n

The old method: How B2B SaaS has approached reviews<\/h2>\n\n\n\n

Having worked in SaaS for over a decade, I\u2019ve seen firsthand how companies, especially startups, solicit reviews. <\/p>\n\n\n\n

In most cases, they rely on incentivization campaigns targeted at happy customers. They\u2019ll also send requests immediately after product milestones or paid conversions, both times at which customers are usually satisfied.<\/p>\n\n\n\n

Here\u2019s how the process tends to play out: <\/p>\n\n\n\n

    \n
  1. Run regular surveys to identify happy customers.<\/li>\n\n\n\n
  2. Send them a direct review request, often from the CEO.<\/li>\n\n\n\n
  3. Also trigger outreach after key product or paid conversion milestones.\u00a0<\/li>\n\n\n\n
  4. Incentivize reviews with vouchers or prize draws.\u00a0<\/li>\n\n\n\n
  5. Add some urgency, e.g., \u201cFor the first 100 people only,\u201d for good measure.\u00a0<\/li>\n<\/ol>\n\n\n\n

    This method has led many to complain about bias and incentivization. Some have described it as a form of soft gating. One startup founder I spoke to, Toni Hoppen of LandingRabbit<\/a>, has chosen not to pursue these highly incentivized reviews at all. I suspect his feeling is more pervasive among founders than is widely acknowledged.\u00a0\u00a0<\/p>\n\n\n\n

    \u201cOn some platforms, reviews are marketing-generated,\u201d he told me. \u201cOn others, the reviews might not represent the whole customer base (just the loudest). \u2026 For AI search, reviews are just one piece of data.\u201d<\/p>\n\n\n\n

    However, despite obvious flaws, the incentivization method has traditionally worked fairly well. One proposed reason is that buyers treat reviews, in large part, like they\u2019re case studies and testimonials. They use them for late-stage evaluation and validation, so their bias isn\u2019t a deal breaker.  <\/p>\n\n\n\n

    According to Axel Lavergne, founder of the review management platform Reviewflowz, people often misunderstand review sites in precisely this way. \u201cG2, Capterra, and any software review platform call themselves review platforms,\u201d he said, \u201cbut they’re really testimonial platforms. Always have been.\u201d<\/p>\n\n\n\n

    \"G2<\/figure>\n\n\n\n

    He was at pains to point out that this isn\u2019t<\/em> a bad thing. \u201cTestimonials are ten times more valuable than anonymous ‘objective’ reviews in the B2B space and arguably in many other industries,” he said, describing the relevance of relatable, detail-heavy testimonials during software evaluation.<\/p>\n\n\n\n

    However, bring LLMs into the mix, and other major shortcomings of the \u201cNPS promoter strategy\u201d become more serious. To understand why, we need to consider two factors: where<\/em> reviews sit in AI-based buyer journeys and how<\/em> users interact with them using AI tools. <\/p>\n\n\n\n

    Where do reviews sit in the B2B buyer journey?<\/h2>\n\n\n\n

    Much and more has been written about the B2B buyer journey<\/a>. One of the issues with popular models is that they assume a linearity that multiple studies have shown doesn\u2019t exist. A study by Dreamdata<\/a>, for example, found the average journey involves 10 stakeholders and spans 88 touchpoints across 4 channels. In addition, 82% of buyers<\/a> enter the market with a preferred product in mind.<\/p>\n\n\n\n

    \"Gartner<\/figure>\n\n\n\n

    For this reason, I like Gartner\u2019s buying-jobs model<\/a>, which understands the buyer journey as a set of tasks in need of completion but not in any pre-set order. It identifies these tasks as problem identification, solution exploration, requirements building, supplier selection, validation, and consensus creation.<\/p>\n\n\n\n

    This model can also help illuminate how reviews act as later-stage evaluation and validation collateral for an existing<\/em> shortlist.<\/p>\n\n\n\n

    Here are the three stages where reviews fit into the model: <\/p>\n\n\n\n