How we build

April 27, 2016

For over 15 years, Dexem has been building SaaS & mobile products used by thousands of people worldwide every day. Customers requests + our vision + unescapable migrations have lead to an impressive set of possible developments, more than we can honnestly handle with to stay focus on delivering the best value to our customers.

Since then, traditional questions about product vision, about roadmap prioritization and about make-or-not-make this user request start to haunt our product management, with the concern to remain focus. Here's how we do to build customer-centric software products.

Building Pyramids

Pyramids are a pur engineering success, from foundations to the very top, which crossed millenia while remaining robust, elegant and fascinating. The parallel with software development can be daring, but I believe it makes sense. With strong foundations, aligned middle stages and perfectly shaped top, software products can be very delightful for users.


We're used to segment our roadmap into 5 main themes which help us prioritize our developments: uptime > bug fixes > performances > features > UX/UI. Very high uptime plus no bugs are our strong foundations, and amazing UX/UI is our very top.

Fixing what decrease customer satisfaction

We make as a - strong - priority everything that could decrease customer satisfaction. This is why uptime and no bugs are our foundations. If you have an amazing UX/UI but are down 50% of time or very slow, well let's be honest, your customer will quickly start to be very unhappy. On the contrary, if you only have a it's ok UX/UI but if are over 99% up plus have very few bugs plus are fast, your customers will get the job-to-be-done by your product.


We want to deliver amazing service quality to our customers so we traget to be over 99,99% on a yearly basis (0,001% a year is actually 9 hours). At the end of April '16, we're 100% for the last 12 months. To do so, we built a reliable servers architecture with automatic replication on back-up servers. And we use Pingdom as monitoring tool.

Bug fixes

Bugs are killing customers because they need + pay to use the product we build to run their job well. When there's a but, they can't do their job well. In order to deploy highest bugs free releases, we invest a lot into automatic testing and now reach a large part of our code with test coverage. When we find a bug and when a customer reports one, we consider it with a high priority and manage our roadmap to fix it as fast as possible.


My take is that performances are often under-rated because development teams have generally fast machines and strong Internet connexion, and it's sometimes hard to put ourselves in place of customers' real life (low connexion, old machine, etc). But it's actually fundamental for customers to get the job-to-be-done by the product with a good speed. To do that, we refine our code performance to load faster and have data centers worldwide to be closer to our customers. And we use New Relic as monitoring tool.

Investing on what increase customer satisfaction

In addition to fixing everything that could decrease customer satisfaction, we do invest in everything that could increase customer satisfaction. We try to bring the features our customers need to get the job-to-be-done by the product they buy from us, with a constant and fluid experience while using it.


Features, features, features... I'm pretty sure your backlog is fully loaded with feature requests. The hard thing are what to make / what to avoid, and how to priorize what to make. We're used to consider every feature within one of the following categories:

  • Game changer features
  • Show stopper features
  • Distraction features
  • Marketing features

Show stoppers are features your customers do really need, you won't sell if you don't have it. For example, a car with no wheels (does that make sense??). Game changers are features that are unique (diffenciation from the competitive landscape) so your customers want to buy your product. For example, an eletric with auto-pilot car (Go Tesla!). Distractions are features you do not need to sell, that some customers will ask for but never use and that actually distort the product from your vision. I know they're filling your backlog but please, consider serioulsy not to bring them in. And at least, Marketing features help you understand how customers use the product (user tracking, features adoption) and help grow customer acquisition automatically.


Last but not least, when having amazing uptime + no bugs + high performances + the right set of features, great design will blow customers' mind. Why? Because it will bring positive emotions to people using the product to get the job-to-be-done. We strongly invest to bring a constant UX with lovely UI, so we delight users to be fan of our products.

Enjoyed the article?

I'm sharing what I build and learn on SaaS products strategy and SaaS distribution. And I'm learning a lot, so will you! Delivered once a month for free.

I won't spam you or sell your address. Unsubscribe - of course - at any time.