This week’s Friday 15, co-hosted by Brian Beck and Jared Blank, took on a problem every B2B operator recognizes: the gap between knowing your data matters and actually trusting it. Brian opened with the numbers that capture the tension. 90% of businesses say high-quality data is essential, but only 43% are confident in their own data quality, and Gartner has found that 40% of businesses fail to hit their objectives because of missing, incomplete, or inaccurate data. As Brian put it, quoting Andrew Ng, data is food for AI. To talk through what that means in practice, the show brought on Bryan House, CEO of Elastic Path.
Why catalog data is the real bottleneck
House’s core point is that in B2B, most of the information a buyer needs has historically lived in the heads of salespeople, in PDFs, in technical specifications, and in back-end systems like the ERP. The salesperson has been the translator and aggregator who finds and assembles all of it. He compared it to the way London cab drivers were once tested on memorizing every street before GPS existed. Once the map is accessible to everyone, knowing where to find things becomes a diminishing source of value. The institutional knowledge has to come out of people’s heads and into structured, accessible data.
Answer engines are becoming the discovery channel
House sees B2C as a leading indicator for B2B. Now that LLMs have established themselves as a discovery channel, B2C sellers are seeing larger baskets and higher conversion, because a buyer who arrives from an answer engine is high intent. They are looking for a specific thing, and they purchase when they find it. That same behavior is moving into B2B, and the risk is straightforward.
The downside risk is that people aren’t finding your products, because they’re finding other solutions instead.
Bryan House, CEO, Elastic Path
The old goal was being on the front page of Google search results. The new goal is making sure the answer engines understand your products and your unique capabilities well enough to recommend you to interested buyers, especially as the buying population gets younger and leans toward self-service.
Where to start without boiling the ocean
Jared asked the question on most practitioners’ minds: is this something an ecommerce leader can own, or does it require a full digital transformation driven from the top? House’s answer leaned practical. The danger is treating it as a boil-the-ocean project. Instead, start with the catalog, assess where the data is good and where information is missing, and take a blocking-and-tackling approach to filling the gaps, whether through AI enrichment or by connecting back-end systems. He is not a fan of big-bang digital transformation. What the effort does need is a leader with a mandate, someone willing to own the initiative and do the internal work of getting stakeholders on board.
What it takes beyond leadership
Two things round out the leadership piece. First, a clear understanding of what your customers actually need to make a decision, much of which already lives inside your sales organization and helps narrow the scope of the work. Second, data sources that are accessible by API, with middleware in place where older systems cannot be read by modern tools. The information has to be machine-readable before an LLM can use it. House also noted a counterintuitive pattern: mid-market companies with homegrown systems sometimes have better data hygiene, because the teams that built their own commerce channel tend to be hands-on owners of their data.
How you know it’s working
The proof shows up in familiar metrics. House’s B2B customers tend to see average order value and conversion rates climb, which builds the business case to expand from one market into others. On top of that, teams are starting to track how much traffic comes from LLMs. It is still a small share, but it is worth watching, because answer-engine traffic tends to convert better than other sources given the intent behind it. Elastic Path is seeing the same dynamic from the other side: a growing amount of its own inbound interest comes from buyers who found the company through an LLM, a virtuous cycle that rewards exactly the kind of data work House recommends.