I’m a business analyst, and a big part of my job involves working with engineers and product managers to gather detailed, in-depth information. For reasons I don’t fully understand (though I have my theories), I often find that engineers, in particular, seem oddly reluctant to share the information I need. This makes the process more challenging than I’d like. Does anyone have tips or tricks for building trust with engineers to encourage them to share information more willingly and quickly?

EDIT: Here’s a summary with more details for those who requested more info: I’m working on optimizing processes related to our in-house file ingestion system, which we’ve been piecing together over time to handle tasks it wasn’t originally designed for. The system works well enough now, but it’s still very much a MacGyver setup—duct tape and dental floss holding things together. We got through crunch time with it, but now the goal is to refine and smooth everything out into a process that’s efficient, clear, and easy for everyone to follow.

Part of this involves getting all the disparate systems and communication silos talking to each other in a unified way—JIRA is going to be the hub for that. My job is to make sure that the entire pipeline—from ticket creation, to file ingestion, to processing and output—is documented thoroughly (but not pedantically) and that all teams involved understand what’s required of them and why.

Where I’m running into challenges is in gathering the nitty-gritty technical details from engineers. I need to understand how their processes work today, how they’ve solved past issues, and what they think would make things better in an ideal world. But I think there’s some hesitation because they’re worried about “incriminating” themselves or having mistakes come back to haunt them.

I’ve tried to make it clear that I’m not interested in punishing anyone for past decisions or mistakes—on the contrary, I want to learn from them to create a better process moving forward. My goal is to collaborate and make their jobs easier, not harder, but I think building trust and comfort will take more time.

If anyone has strategies for improving communication with engineers—especially around getting them to open up about technical details without fear—I am all ears.

  • @FourPacketsOfPeanuts
    9•14 hours ago

    They are probably unsure of your motives; are you analysing the business or analysing them? Software problems are extremely hard to estimate unless there is almost complete disclosure and discovery. It’s like asking people how long a crossword is going to take without seeing the clues. Or asking how long they’re going to spend on a chess move in 3 turn’s time. They are possibly cagey because you are asking questions that betray the fact you are seeing this as a management problem rather than listening to what they’re telling you about their craft.

    Or possibly your manner of communicating is attuned to more socially intuitive people. Try presenting what you need as a problem for them to solve with a clear start and end. That way you’re collaborating, and they know when their obligation to interact with you is “done”.

    Instead of open questions like “can you tell me how X is currently working?” try specific problem setting questions like “I’d like to see if we can make X process be 10% faster, what would that look like?” or “what would you say are the top two things that affect the time process Y takes?”

    They may not want to offend you, because many of the answers might be “obvious” and, also, if they’re honest workers, as many are, there may not be any clear way to improve certain things as they’re already trying their hardest, and your investigation feels more like an inquisition.

    Again, it may be that you’re asking someone “how can I get you to get this crossword done faster?”. It’s sort of the wrong question. Unless you’re willing to listen to their bugbears which might be the actual things affecting how efficiently things run but might not be the kind of answers project management want to hear.

    • DominusOfMegadeusOP
      3•15 hours ago

      I’m at the stage of “I want to know how your process has worked up until now, and how you would like it to work, in a perfect world.” Which did seem to garner a positive response.

      • @laurathepluralized
        3•3 hours ago

        To add onto this, sometimes it’s about getting more specific with your questions to get the more specific answers.

        For context of how I would suggest structuring these detail questions, here’s how I think about code I write or debug: The functions and classes my code is made of are meant to get specific inputs to become specific outputs via a defined process; I think of this as inputs->how->outputs. Figuring out what inputs you need to execute the “how” part to get the outputs you want is the puzzle of each function or class I write. The “how” part can even be broken down further into smaller chunks of inputs->how->outputs.

        I think asking your engineer friends to frame things in this context would both show your appreciation for the nitty-gritty details you are needing, as well as give you further context to ask more detailed drill-down questions (about deeper levels of inputs->how-> outputs) if needed. For example: “you said to get inputs A and B to result in output C, we need to run the fizzbuzz algorithm on A and B. What roles do those inputs have in that algo? Do we have to do any preprocessing on A or B before we fizzbuzz them, or any post processing of the fizzbuzz’s direct output to get C?” “Oh, yeah, we have a wrapper that takes A and makes it column-major so that fizzbuzz executes faster, but we need output C to be row-major for when it goes into otherFunction(), so we do such-and-such to fizzbuzz’s output to get the C we output.” This gets you a level of detail deeper, and you could ask further questions about the transformations happening to A and the post-processing of fizzbuzz output to get C, as well as get more context for otherFunction to ask more about later.

        You could also use this context to ask further questions about what they think the future implementation should look like. “Are there any assumptions we can make about A and B or how C is used that could simplify how we go from the former to the latter? Are there any requirements on the inputs and outputs that would better be either relaxed or made more stringent, and if so, in what way?”

        I hope that helps! Best wishes for your work on this project–streamlining processes is hard, especially when working with other people’s code, but your appreciation for the details to get things implemented well is admirable!