Facilitating the Spread of Knowledge and Innovation in Professional Software Development

Write for InfoQ


Choose your language

InfoQ Homepage News Innovation at AXA's Digital Agency

Innovation at AXA's Digital Agency

This item in japanese

Lire ce contenu en français

The AXA Digital Agency deploys the Lean Startup approach, using design thinking, minimum viable product development and growth hacking, to innovate and support the digital transformation at AXA.

At the Lean IT Summit 2015 Yves Caseau, Head of the Digital Agency of the AXA Group, will give a talk about Lean Startup as the innovation engine for the Digital Agency at AXA.

The Lean IT Summit will be held in Paris, France, on October 8 -9, 2015. InfoQ will be covering the event with news, Q&As and articles.

InfoQ did an interview with Caseau about the importance of innovation for AXA, adopting a lean startup approach for innovation, learnings from using minimum viable products and growth hacking and advice for starting an innovation journey.

InfoQ: Can you explain how important innovation is for AXA?

Caseau: Innovation is about bringing value to consumers by solving "pain points" that they have, even if they are not necessarily aware of them beforehand. Innovation is necessary to most companies to sustain customer satisfaction but is is especially important for companies such as AXA that are ongoing a massive digital transformation. I am the Group Head of Digital at AXA whose job is to develop new "digital experiences", mostly though mobile applications, that feed and demonstrate AXA’s ambition to protect our customers.

InfoQ: What made you decide to adopt a lean startup approach for innovation?

Caseau: Lean Startup is, in general, a great way to develop innovation in the digital world. I have become familiar with the Lean Startup approach during my previous job at Bouygues Telecom, when I was in charge of developing new services and new products, such as Internet gateways and Set-top boxes.

I have brought Lean Startup to AXA because it has two key features. First, it is an incremental "product centric" method (as opposed to project centric) that helps to work with agile development methods. A traditional, cultural approach in an insurance company to to get "everything right on the first time". Digital state-of-the-art is the result of continuous improvement with many iterations. Second, Lean Startup is a customer centric approach, where a significant part of the value is co-developed with users. This is why Lean Startup advocates for building a minimum viable product (MVP) rapidly, to start this meaningful dialog with users as soon as possible. Learning from customers requires the support of the product: it is a practical discusion about usage, not a theoretical one about what would be nice or useful.

InfoQ: Can you share how you view innovation?

Caseau: Innovation at AXA Digital Agency follows three steps which are part of the Lean Startup approach : Design Thinking, MVP development and Growth Hacking. The first step is really about pain points : finding them, understanding them, creating "User Value Propositions" (UVP). Our methodology matches Ash Maurya’s "Running Lean" pretty closely. Discussing with and listening to customers starts early, both to identify problems ("pain storming") and to creatively think about possible solutions. The whole issue is to design a complete "experience" that encompasses products and services, and which is completely centered around the future customer. We have started to develop "emotion design" approaches, but this is a long journey, we are still learning! The second step is to build the first version of the mobile application. Prototyping has started earlier, as soon as we start design, by the way. In the second step, we aim to build a true "viable product" using state-of-the-art agile development methods. By state-of-the-art, I mean that we use DevOps, continuous build and delivery, automated testing ... it is also a long journey and we have not reached the status of what I would call a "Lean Software Factory", as described in my previous Lean IT Summit talk, but we are making progress. A key dimension of the Lean Startup is to "measure everything", because the first goal of the MVP is to learn from users. Thus, we make abundant use of "mobile analytics".

InfoQ: Can you give some examples how you have used minimum viable products? What have you learned from using them?

Caseau: "Minimum Viable Product" is a falsely simple concept, since it is a true oxymoron. By "minimal", we aim to implement as few features as possible, but with enough value to solve a problem and to deliver value. In a MVP, the goal is to get there fast, because we know that we will only really learn once we have a product on the market. But it needs to be a "viable" product, that is a product that brings value. It is often said that Lean Startup implements the Silicon Valley saying "Fail fast to succeed faster". But if the product is not great, failure is garanteed ! Nathan Furr and Jeff Deyer used the term "Minimum Awesome Product" to carry this point : the MVP needs to be an awesome product that delivers the promise that was captured with the UVP (Unique Value Proposition). This MVP paradigm is very well suited to mobile application development. The Digital Agency builds the first version of its smartphone applications, such as AXA Drive or AXA Health, as MVPs. We work hard to make sure that the first version delivers value, although we know that we will learn much more in the future from actual usage of real users, which will feed the development of subsequent versions.

InfoQ: AXA uses growth hacking in their marketing approach. Can you briefly describe what it is and how you use it?

Caseau: Growth hacking is the third step of mobile application development for AXA’s Digital Agency. It includes improving the application and growing the usage, both for each user individually and as a community. We have coin an expression to represent this continuous learning loop from our users : CFLL (Customer Feedback Learning Loop). We learn both implicitly though analytics (cf. previous comment: we measure everything) and explicitly (using many types of feedback forms), by "talking to users". We also develop a "social community" of application users, to harvest the social dynamics of feedback (people tell you much more about your application if their feedback is valued within a community). Growth hacking is actually about developing the size of your regular users community, not only those who want to talk about it. Growth Hacking, like Lean Startup, is a growing trend with a large number of practitioners. There are a lot of techniques to be learned, to develop the viral adoption of your digital innovation. A key idea is to make your product "its own distribution channel". Growth hacking comes from the combination of marketing and code development, precisely because digital technology makes this combination meaningful and fruitful.

InfoQ: What are the main things that you have learned on your innovation journey. Any advice that you want to give to the InfoQ readers?

Caseau: What I really like about the Lean Startup approach and the many books by Eric Ries, Ash Maurya or Nathan Furr is that they explain very well the many mistakes that many innovators, including myself, have done in the past. Part of the reasons why I am such a big fan come from the poorly designed services or applications that I have contributed to in the past, that resulted in disappointing results. It is hard to capture this experience in a few sentences, but my main advice is to read these books and the abundant web literature about Lean Startup. Then, just as is the case with lean, this is not something that you learn by reading, but by doing. So my second advice is to select a small problem and use it as a toy to develop a complete "lean startup" agile development experiment. The main lesson from "Running Lean" and many similar books is that we tend to move too fast to development versus problem characterization. This is another paradox: we aim to develop a small MVP as fast as possible, but it is essential to spend as much time as possible to characterize the "job to be done", the "problem that matters". My last advice would indeed to make sure that you are not developing a product that nobody wants, because this is something that happens too often.

Rate this Article