The customer who reaches out when they have to

The signal is already in the data

B&H’s B2B contact center is inbound-only. Customers call or email across the full transaction lifecycle, from pre-order questions and quotes through order status and post-purchase support. Shmuel has spent 12 years learning to read customers who are frustrated. What’s harder to read is the one whose contact frequency has quietly dropped from daily to weekly, whose order volume has ticked down with no complaint filed. That customer isn’t upset. They’re just less present.

The current feedback system is structured. A dedicated survey platform reaches customers and a team categorizes every response into monthly and quarterly reports. But as Shmuel put it, surveys are a moment in time. By the time you have the data, the customer was already upset. The customers most at risk are the ones who send nothing back. They may just buy less next time.

The data that could answer this is already in the system. Order history and contact patterns hold the signal: a customer who called daily and is now calling once a week. Brett raised it concretely during the conversation, walking through a repeat customer whose checkout sessions gradually lengthened and then abandoned on the fifth visit. That behavioral drift is exactly what Shmuel is trying to act on earlier. AI tools that model behavior and test scenarios before pushing to production are part of where he thinks this goes. The harder question is who owns that signal when every department is looking at the same customer through a different window.

The signals are in the data, and the question is how do you read them early enough, reliably enough, to actually do something about it?

What Shmuel’s teams actually built

Shmuel oversees four inbound B2B contact center teams at B&H Photo and Video. Each team is run by an account manager with eight to nine people alongside them. The operation covers government clients at the local, state, and federal level and education clients from K-12 through higher ed, mostly domestic with some international work. Every interaction is inbound. The outbound relationship work, visiting accounts and attending trade shows, lives in a separate part of the company.

For email, the team runs Salesforce. Phone calls run through Genesys. A homegrown transactional system handles the operational work specific to B&H’s ordering process. Chat is available on the consumer side of the business but not in the B2B operation. Customer feedback comes through a dedicated survey platform; a team categorizes every response into monthly and quarterly reports that go to leadership.

The recognition system Shmuel built is called Feel Good Fridays. Each week he pulls a specific positive interaction per team from an email thread or a survey comment and names the person in a shared channel that spans his entire division. Then he opens the floor for everyone else to do the same. He’s methodical about it: his goal is to find the ordinary, well-executed interaction, the regular job done right. It could be a teammate who stayed on a call a few extra minutes so someone else’s lunch break didn’t run over, or notes left on an order that made the next handoff cleaner.

Where Shmuel is candid about the gaps: the survey platform captures what customers choose to say, and a substantial portion don’t say anything. Those customers are harder to quantify and, in his read, more at risk. The feedback also lives across departments in ways that don’t connect. Returns, product, logistics, and the contact center are all talking to the same customer. None of them share a unified view, and who owns the work of consolidating that feedback is still open.

One thread that came up later in the conversation: as AI handles lower-complexity work, what’s left in the queue is emotionally harder. Shmuel’s already thinking about what that shift means for his people. If a password reset was a mental break between more demanding calls, removing those interactions changes the rhythm of the day. He’s reconsidering how he structures schedules and reads metrics because of it. Handle time going up on a complex call doesn’t mean something went wrong.

Shmuel hasn’t solved the silent customer problem yet. He knows the signal is somewhere in the behavioral data and has a sense of where the tools are heading. What he’s still working on is how to surface it early enough to act on, and who in the organization picks up the work of making that data operational before the customer is already gone.

If you’re running a similar inbound B2B operation and you’ve made progress on this, Shmuel’s on LinkedIn.

• Shmuel Saklad’s LinkedIn
• B&H Photo and Video
• Salesforce
• Genesys

• Youtube
• Spotify