Turning B2B Search Into a Tenfold ROI on SAP Commerce Cloud

When Deirdre Peters joined Boston Scientific in 2019 as its first e-commerce hire, the company’s product catalog was still printed on paper. SAPinsider covered that starting point and the headline results in August. A recent SAPinsider webinar added the details behind the numbers.

Peters, Director of Digital Experience at Boston Scientific, joined Peter Curran, Chief Product Officer at Coveo, and Winnie Albornoz, Senior Director of Product Marketing at SAP, to explain how a medical device company with 12,000 products, 127 countries, and more than $20 billion in annual sales turned product search into its first AI investment with a measurable return.

The Problem Was Not the Platform. It Was Finding Things.

Boston Scientific launched e-commerce in 2022 on SAP Commerce Cloud, chosen for its B2B depth, global scalability, and built-in PIM. Adoption was strong, driven by the 10,000-person field sales force. Then the feedback shifted. “Your site is awesome. It saves me so much time, but I’m having trouble finding products on it,” Peters recalled hearing. Search conversion fell, and so did the number of customers who searched.

Buyer behavior also changed. “People are no longer searching for product numbers. They are searching with questions,” Peters said. “Does this product work with another one? Is it compatible with MRI machines?”

Albornoz sees the same pattern across SAP’s B2B customer base. “When you’re a B2B buyer, you use search to answer questions. You have your business in mind,” she said. The expectation has hardened over time. “Many people were saying that B2B buyers want a B2C experience, but now it’s expected. It’s not just a want, it’s a need.”

Where the Platform Ends and the Partner Begins

Albornoz was direct about the limits of native search on SAP Commerce Cloud and why Coveo is an SAP-endorsed partner. “We do provide search capabilities, but it really depends on your digital maturity,” she said. “When Boston Scientific started off digitizing the catalog, the search was good enough. And when it came time that you needed more complexity, Coveo came in at that point and met that need.”

Curran said the relationship pushes Coveo to stay ahead of what the most advanced SAP customers need, including extending SAP’s own agents “so that we give them a kit of capabilities that they wouldn’t have out of the box.”

An AI Use Case With a Business Case Attached

Peters also faced a familiar question from the C-suite: what is the company doing in AI? Coveo offered a way to answer that with a measurable return rather than an experiment.

Working with Coveo’s business value team, she modeled the impact of restoring search engagement and improving relevance. “I presented a pretty aggressive use case internally, a business case, that promised about a million dollars of return on a couple hundred thousand dollar investment in the first year,” Peters said. “We saw almost tenfold that result.”

Curran explained why Coveo treats that pre-sale modeling as more than a sales tool. The same business value team returns after go-live to compare results against the original projection. “When we have that same business value team able to look before and then look after,” he said, “that really helps focus the effort of the tech team to get things into shape and realize the ROI that we projected.”

The Results

Boston Scientific reported a 24% increase in search usage, an 18% lift in search conversion, 16% incremental e-commerce revenue growth over any other channel, and a $6 million search revenue lift in a single measurement period.

Availability-aware results surface approved alternatives when an item is on backorder, a direct revenue lever for reps in the field. The lean merchandising team no longer maintains synonym lists and conflicting rules.

Do Not Wait for Clean Data

Asked about enterprise AI adoption patterns, Curran challenged a common sequencing assumption. “You don’t get to done in data cleanup. It’s an evergreen problem,” he said. “The idea that you can get to done and then you start the AI project is a fallacy. You start them both.” Real queries, such as a physician asking for “the forceps with an orange box,” reveal which product attributes are missing from the record.

Albornoz framed the risk of skipping the foundation entirely. “There was a stat that I saw recently from Gartner that said 40% of AI projects by 2027 will be scrapped because there is no architectural foundation or governance,” she said. “Everyone wants to do AI, but you can’t have AI to scale without helpful data.”

What This Means for SAPinsiders

• Search is the highest-leverage AI entry point on SAP Commerce Cloud. It produces measurable conversion and revenue outcomes, making it easier to fund and defend than open-ended AI pilots.
• Build the business case with the vendor, then hold both sides to it. Boston Scientific’s before-and-after value tracking turned a projection into a reinvestment engine for personalization and global rollout.
• Start AI and data governance in parallel. Live customer queries expose which product attributes matter. Waiting for a finished PIM cleanup delays value and risks fixing the wrong data.

Watch the full webinar. Peters, Curran, and Albornoz cover the sales council that drove rep adoption, why change management was 90% of the job, and how Boston Scientific is preparing for agentic commerce. Watch “How Boston Scientific is Reimagining B2B Product Discovery” on demand.