Last week I gave a private talk to a large international pharma company's analytics/BI/DS team and we had a great team discussion afterwards for about an hour. One of the things that came up in this was a common challenge I've seen at really large companies: getting access to the actual end users (not some proxy person). The concern? Too many gatekeepers and politics, or simply the fact that too many people (with good intentions) are all trying to access the same people.

Now, this is actually a nice problem to have—it assumes that your team really is invested in building innovative ML and analytics solutions. They WANT to talk to customers and see how their work fits in. They have already made the decision that design and UX matter. That is something to celebrate, because in a lot of places, the culture is more like this:

"I have no idea what they're going to do with this data, but I'm doing what I'm told—here's the dashboard/excel you requested."  Next Jira ticket please.

THat's not how you create innovative ML and analytics products.

If you're in the former camp where the problem of NOT talking to customers has already been recognized and people are ready to get out there and learn, then let's look at the next challenge: access.

Why access to real users / customers is blocked at some enterprise companies

As companies scale into the 1000's of employees, it's only normal for politics to creep in, for the status quo to rule and for change to be inhibited.

Leaders need to be aware of this, and step in to make changes—including clearing the space for what I think is the #1 most important activity in the process of designing innovative ML and analytics products: 1x1 customer research with the people who will use the service/solution/apps. It's not playing with data, building things, coding things, or drawing things.

Access is not always a trivial matter of just emailing or calling somebody. Simply figuring out WHO is a customer can be difficult. You may have sales people, sales engineers, customer service people, market researchers, business process people, and a whole host of other roles potentially trying to talk to the "users" as well—or prevent your team from doing so at first.

Typically, for data teams serving internal users, there is much less excuse and it's a lot easier to tap your customer—they're getting paid to support the same mission as you, after all. When your user is a paying/end customer though, this needs to be handled differently sometimes—but not to the point that the org. creates a bureaucracy that effectively prevents any meaningful contact between the creators (designers-i.e. YOU) and these users.

So, what to do? Let's look at some options.

Getting access to customers so you can do 1x1 qualitative research

First, as a leader, you have to develop relationships with the existing stakeholders and gatekeepers to your real end-users/customers. There may already be a centralized customer-research group to manage all the "touches" and a line of colleagues in your company who also want direct customer access. Understand that some of these people, such as sales people, may be scared to let you access customers—for fear you may promise the wrong then, scare them away, or jinx a sale/relationship.

Other people may be (rightfully) concerned about being too much of a "tax" on a specific customer and will want to do the proper internal research to make sure that your company isn't going back to the buffet for too many extra helpings.

These are all valid concerns—just don't let the bureaucracy get in the way. And if there isn't a department, that's good in some ways - in the short term, you'll be able to make some easier customer relationships and develop what we sometimes call "design partners"—customers who are interested in being long-term resources for your team as you work on solutions. Aka "friendlies."

Handling gatekeepers and "competing" departments

When there is a gatekeeper already "in the way" of your team's customer research, try to understand the gatekeeper's process, how they grant access to customers, and their concerns. Find ways to potentially combine research activities—i.e. you may be able to combine two research sessions into one with another department who also needs "face time." You may find out that the information each of you gets benefits each other—because your perspectives are different. You can also use the very same empathy and techniques you'll use on your customers with these gatekeepers: figure out what their concerns are, and see if you can help them while you help yourself. Sharing your findings, and finding opportunities to inform these gatekeepers of any relevant insights you get during your team could be really helpful.

Ex: "Hey you know what? Paulette, that director of IT we spoke to mentioned XYZ about next quarter's belt-tightening when we were showing her our dashboard and getting feedback on how she uses it in her work. Something about how all company purchasing may be on hold due to Covid for another 6 months - even with the vaccines coming out. Thought you might want to know we overheard that."

A salesperson would find that feedback VERY useful most likely.

Another way to get access to precious customers?

Find out if the "known" customers have "colleagues" who do the same work/job.

This may be easier than you think.

Getting access to new customers…from customers

For example, many sales leaders are going to be focused on getting access to "decision makers"—i.e. Directors, VPs, SVPs, etc—who are often NOT the people who are going to be living with the technology you may be creating.

So, use that to your advantage: let them know you don't need to talk to a decision maker—you want to talk to the direct end users under them. You can also position this as something that could benefit the salesperson: by going out and talking to the subordinates, you're creating "face time" with the [company] customer. Your org's brand, products, and services are getting more attention—even if you are genuinely there to do research and inform the data products you're creating—not to sell anything. That can be a win-win. I've seen it happen more than once: research sessions end up "warming up" a customer or even making them feel like they are being "trained" or getting special "face time." This isn't always true: sometimes people are really busy, and your cash incentive for participating may be what drives a user to say yes.

Still blocked? It might be time to escalate it to a senior executive who values customer feedback

Another idea if the gatekeepers are just blocking all meaningful access? Escalate this to a senior stakeholder who can break the ice. The thing is, chances are, there is something about "customer centricity" somewhere in the company's long-term goals. At a high enough level, there is almost CERTAINLY an executive leader who very much would not like to hear that her departments are blocking your access to customers when you're on a mission to serve them. Sometimes a casual mention of this in the lunch room can go a long way. "Hey can I ask you something? Why is it so hard for us to get access to real customers so we can do our job? Can you help?" I've seen an excel spreadsheet appear within 24 hours before using this technique, listing out a bunch of customers to reach out. Even better? Sometimes these senior stakeholders will know EXACTLY which types of people to talk to. They may even want to hear what you learn.

Being persistent, but careful, when there is an existing customer-access process in place

On the other hand, I think some caution is warranted if you have many departments all trying to tap a limited number of qualified end users for research inquiries. Some stakeholders will be very concerned about how many times the company is interfacing with customers.

Be sure you understand the last time that customer was spoken to, perhaps ask the employee who last spoke to them to understand any sensitive issues that may exist, and be human. I am sure you've been on the phone with customer service before and they're just like, "Sorry, I don't know: that's another department. We don't have a way to contact them internally. But here's a phone number." Ick. Well guess what? Now you're the equivalent of that CSR rep/person on the end of the phone representing the company. Understand that the customer may need/want something not relevant to your particular work. Take down their concerns, follow up, and make sure it does't get dropped—they'll likely remember, and those gatekeepers will be glad that you didn't forget that your [selfish UX research!] needs aren't the only thing the matters.

I know this may seem like a lot, but hey, if you want to design innovative ML and analytics solutions—particularly external-customer-facing ones.

Or, you can keep guessing and praying.

Free help may be a phone call away—from your own UX dept.

If your company is quite large, your BI/analytics/data team may not be alone, and there may be some "free help" you can tap so you can focus on the data products you're working on and the relevant research questions. See if your organization already has a group/dept. dedicated to handling customer research inquiries. It might. And secondly, even if you haven't traditionally worked with your UX/Design department, they may have a LOT of insight about how to do the type of research you're doing successfully. They may also know how to navigate your organization/company politics, and how to maximize the time spent with the customers. They'll likely be thrilled to know somebody else even cares enough to do qualitative research and will want you to be successful.

As to what to ask customers? Here are some ideas in a previous article I wrote.

Now, get out there, turn your ears up to 11, and start hunting for the unarticulated problems and needs that you can only get from these 1x1 interactions. If you want to design innovative ML and analytics solutions, this is a big part of doing that well in a consistent, repeatable way.

Need more specialized help specific to data products? Here's how to work with me.

Photo by Yann Allegre on Unsplash