The ill-fated Pontiac Aztek: an object lesson of product design process failure.*

The Science of Customer-Led Software Design (Part I)

Note: this is the first part of a discussion of product management; The Art of Customer-Led Software Design is the 2nd installment.


Gain Compliance is trying to solve a complicated business problem.

The attributes of our target market — its complexity and the investment necessary to create a solution — are the hallmark of rich business opportunity.

At the same time, they also create a very real, here-and-now challenge for a startup. We don’t have unlimited time or resources; we need to get working software in customer hands as quickly as possible. From there, we will improve, refine, and expand the product.

Deciding what features to build, in what order, and what they look like — this is the puzzle.

In our agile approach, we base 100% of product strategy and design decisions on customer discovery and feedback. This requires discipline: diligent discovery with a broad range of target customers is a time-consuming, exhaustive, and often plodding affair. It would be far more expedient to take shortcuts.

Good product managers avoid the pitfalls which lead to answers which are easier and faster, but much less reliable, such as:

  • Leading the witness. The best data comes from open-ended conversations, not from questions which suggest a correct answer. An example of this is asking a customer to pass judgment on a design idea or confirm the steps in a process. At the other end of the spectrum, the ideal opportunity is to observe customers as they tackle business challenges in their current state.
  • Trusting your gut. This one is often a topic of debate. Yes, we all know that Apple would never have invented the iPhone had Steve Jobs relied solely on customer feedback. At the same time, none of us at Gain Compliance is Steve Jobs, and our vision for better enterprise software is more iterative than revolutionary. We don’t know better than our customers, and we need to validate significant features and approaches.
  • Seeking approval. As we become invested in our approach and the work we’ve put into it, it’s natural to be proud of our ideas and designs. Our pilot companies are full of nice people, and they will be naturally supportive when providing feedback. These two factors can generate a positive reaction, even when the reality is otherwise. The best process focuses on creating an environment where the feedback is candid.
  • Benchmarking on competitive products. We don’t base feature decisions or software design on what other products do, but rather on what ours should do ideally. Quite simply, we ignore the features and functions of the competitors’ offering and how others have previously solved customer challenges.

What are the steps required to build an attractive product which balances time-to-market with valuable functionality?

First and foremost, the product design role requires someone who has an otherworldly level of discipline, an incredible reserve of patience, and a genuine intellectual curiosity.

Secondly, sufficient runway (time) and resources (talented software engineering team) are needed.

Third, a significant sample of pilot companies must be generous with their time and expertise.

Finally, the product team must stay true to the process and avoid the pitfalls. Listen to the customers — they’ll tell you what to build.


This framework is the science of product management, and it’ll get you pretty far. But building a truly great product requires a little art as well. For a discussion of building software that offers a visionary solution, check out the second installment to this post.

*For an excellent article on how product development can go off the rails — as well as an explanation of how the Pontiac Aztek, widely recognized as the greatest automotive product failure in recent history, came to be — check out this article.

Make sure to check out our other the blog posts, and follow Gain Compliance on LinkedIn, Facebook and Twitter.