They Don’t Shop by Company Size

Overview

I led the user research on the Business Internet shopping experience across several rounds over roughly a year and a half. The work started from a simple observation that kept getting stronger with every study: business buyers do not choose internet the way the product assumed they did. The single most important outcome was reframing how the team thought about organizing the buying experience, moving away from segmenting people by company size and toward organizing around what people actually need. This is a story about holding a research through-line across many rounds and letting the evidence move a decision that was bigger than any one study.

The challenge

When a business owner or a procurement lead shops for internet, they are trying to answer a few very practical questions. Is this available at my address. What will it really cost. Is it reliable enough that my business will not lose money when it goes down. Which option is right for me, and can I trust that I picked correctly. The experience I inherited made people work harder than they should have to answer those questions, and it opened by asking them to sort themselves by company size, which turned out to be the wrong first question. My job was to find out how these buyers actually make decisions, and to make sure every design change was grounded in that reality rather than in how the business was organized internally.

My approach

I treated this as a program rather than a single project, because a page like this cannot be understood in one round. I ran moderated qualitative research, and I built each study on what the last one taught me. I began by analyzing real prospect conversations to learn what people ask before anyone has designed anything, which gave me an honest baseline of the questions buyers bring with them. I then ran moderated testing on the existing experience to see where it fell short, followed by rounds of concept and usability testing on redesign directions. Along the way I added a focused study on how buyers evaluate hardware and a competitive review of how other providers handle the same journey, both of which sharpened the thinking that fed the later rounds.

The most recent round was a comparative concept test with fifteen business prospects, recruited as an even mix across micro, small, mid-market, and enterprise, and across single-location and multi-location businesses. Each participant saw two design directions that took different structural approaches to the same problem, and I counterbalanced the order so about half saw one first and half saw the other, which let me watch for order effects rather than baking one in. I was also deliberate about what the study could and could not fairly compare, so when one reference point was scoped too differently to sit in a clean head-to-head, I said so and reported it as context rather than pretending the comparison was valid. Being clear about the limits of a comparison is part of keeping findings honest.

What I learned about these buyers

A few insights held up across rounds and shaped the product direction.

Business buyers reason by need, not by company size. When people were free to navigate by what they were trying to accomplish, they did that naturally, regardless of how big or small their business was, because an enterprise can want one simple location and a small business can carry needs that look enterprise-level. That single pattern is what moved the whole segmentation model from size-based to needs-based.

Reliability mattered more than speed or price for nearly every buyer, because downtime costs a business money directly, and people evaluated their options through that lens even when the experience did not put reliability front and center.

The information people most needed was often the hardest to find. When the one detail that would let someone decide is buried where they will not look, the experience cannot do its job, and confidence collapses no matter how good the rest of the page is.

Buyers wanted to be guided. Several people, without any prompting, described wanting to answer a few short questions about their business and get a recommendation they could then adjust and control, which told me this was a real and recurring need rather than a one-off request.

Technical detail had to come with plain-language explanations. Tech literacy varied a lot from one buyer to the next, and people without technical staff still had to make the decision themselves, so specifications only helped when paired with a clear explanation of what they meant for a business. This is where my commitment to accessibility showed up in practice, because an experience that only works for the most technical buyer excludes a large share of the people it is meant to serve.

Impact

The through-line I care most about is that the research changed a foundational product assumption. The experience used to ask people to classify themselves by company size, and round after round showed that need does not follow size, which is why the direction moved to organizing the experience around what buyers are trying to do. The most recent round gave the team a clear direction for the next iteration along with a specific new idea to build: the guided, needs-based recommendation experience that buyers had asked for on their own. I delivered the findings and section-level recommendations to a cross-functional team of design, content, and marketing partners, framed to inform their decisions rather than to dictate them.

How I practice

I hold a through-line across many rounds, so that the research questions, the study design, the sessions, the analysis, and the readout all map to each other, and I notice when they start to drift. I keep every finding traceable to what people actually said and did, and I calibrate my language to match, so a claim about all fifteen participants and a claim about a single voice never get the same weight.