Deep dive in a lean digital company #6

Today I am joining Benoît on a project gemba walk. He is doing two of these every week (each sale in this digital world turns into a project, to design and build the app sold to the customer) and he also attends a kaizen debrief at least once a week. When we enter the open space, the project team is waiting for us. A beamer displays on the screen the virtual project board with milestones and KPIs. With many in the team working remotely due to the pandemic, all the visual management that before Covid was manually updated on walls and paperboards had to be replaced by digital boards, shared by the team members from home.

Matthieu, the technical team leader and architect, is starting to comment the project board but Benoît has another idea on how to start the discussion: “Matthieu, can you show us the product first?”

A true lean question. Many still believe lean is all about honed, standardized processes, with a good dose of control at each key step or milestone. Benoît will later tell me that at first they had tried to solve their quality and delivery issues by focusing on processes but found that centrally-designed one-size-fits-all processes deterred people from thinking. The purpose of a gemba walk focused on a project, from Benoît’s point of view, is not to check whether the board is correctly updated or the milestones met, but whether the product under construction will deliver the value sold to the customer.

It turns out that the product is a mobile marketplace app designed to support on-line bids for an iconic auction house in France. Matthieu shows us a sober mobile app where artefacts on sale are listed together with a picture and description. He surfs the app to show the features already built, such as the details for the upcoming sales (Benoît will ask him to show where he clicks as he wants to understand the navigation on the app from the user point of view). At this stage (end July), the app has not yet been released to production: mock releases are scheduled for the end August (on users’ registration) and user testing is planned for the end of September.

Ghislain, the part-time Project Director, explains the context: “They have a web app that is a bit outdated. New 100% digital auction actors have stepped in over the last 10 years, and their historical competitors – such as Sotheby’s or Christie’s – have already developed on-line bids capability. Our customer needs to catch up, or even gain a competitive edge. Hence the idea of a mobile app (rather than a web page) that would deliver multiple features with a wow effect.”

Benoît challenges the team. “Multiple features and a wow… Are you sure about that?” The project team is present but so is the C-suite of BAM, one of the Theodo spin-offs, in which the gemba is taking place. Guillaume, the head of the Flutter Tribe (Flutter is the technology with which the app is coded), nods: “You mean there is a risk that too many features will never produce a wow factor?” Clothilde, the product owner, confirms that painful slashing on desired features has already taken place: “Each stakeholder on the customer side had a story to tell that seemed to justify a feature.”

Gemba walks are always opportunities to learn on the job, and this is no exception. On-line bids are asynchronous and come in over a pre-defined period of time. Traditional auctions, on the other hand, take place on a specific date and time, when everyone bids simultaneously, either in the auction room or through a remote device. As one of the digital competitors has already announced the release of an on-line bid app, BAM’s customer aims to develop a mobile app that would both cater for 1) remote bidding in a traditional auctions and 2) on-line asynchronous bids.

Clothilde shows the four success key points agreed upon with the customer, which range from increased sales or enhanced registration to bids to the development of mobile bidders and better customer satisfaction so as to boost loyalty. Benoît nods but enquires further: “I see a lot of what I would call outcome KPIs. But how do you measure the factors that will lead to this outcome?”

Clothilde has some answers as she has listed end-user expectations – such as feeling confident in the bid’s process and the auctioneer expertise, easy registration, smooth browsing of the artefacts, and clear signals on who won the bid among others. Some of these can be easily measured (such as the time to control a bidder’s ID), but not all of them – she admits.


WORRYING ABOUT THE TIME-TO-MARKET

During the meeting, Benoît is clearly using the TPS (Toyota Production System) as a framework. He is now back on the schedule: “Can’t we test some transactions before the end of August? We have four weeks ahead of us.” The team looks uncertain and Marek, the CTO, steps in: “We started the build mid-June and the only thing we can test by the end of August is the registration steps. Couldn’t we have done it faster?”

The advantage of using the TPS as a gemba framework is that it naturally leads to the right questions.  Next to, and part of, customer satisfaction is an obsession with the lead-time. Clothilde, as product owner, explains: “We discovered rather late that the bidder ID verification was the pain point of the current web page. It takes up to two days today, resulting in a large number of no-shows at the auction. We launched a benchmark of the market and our customer is now contracting with a new partner that offers an ID control within a few minutes. But the whole process took time.”

Benoît steps in: “How can we learn from that? Do we have a standard that could help us spot this kind of issue sooner?”

Baptiste, BAM’s CEO, and Marek, the CTO, have an idea on what could be improved. “It may be linked to the project staffing. Profiles with the right know-how may come in too late in the project to spot this kind of issue.”


JIDOKA – SPOT ISSUES AND SOLVE THEM

Benoît has a tool in mind that can help detect problems as they occur. “Can we see the weekly customer feedback?” he asks. If you have read the previous article, you will know that every project team sends a weekly questionnaire to their customer to test their satisfaction with velocity and support.

Clothilde and Matthieu display those on the screen. On the last July sprint, the customer has given a 3 mark – below expectation. Matthieu confirms: “One batch of customer tickets in the database was not taken into account in the sprint. Beyond that, I confirm we have a productivity problem, we do not close our tickets fast enough. We are one week late on the first batch of user stories.”

I see no frowning, hear no sighing, and witness no body language indicating unease or exasperation. There is now a long tradition in the Theodo Group of admitting to problems and discussing them openly. “We have opened a problem-solving form on our productivity, but not yet on this undetected batch. We could not have absorbed it anyway,” says Matthieu. The team sets time aside for problem solving twice a week and they show us a number of problem-solving forms on speed or quality. Benoît does not comment the content but has some questions, most of them directly or indirectly aimed at management: “Were you trained on problem solving, Matthieu?” “Is anybody helping you on that?” “Productivity in projects is a known issue, is there a specific know-how here in BAM on this typical issue?” The answers will confirm the need to train on problem solving and better share know-how on standard issues.

Benoît is back on the overall lead-time. “It seems you have a delivery issue. Do you maintain a macro-schedule on your deliverables?” he asks.

Matthieu has developed a schedule on Excel, with the downside that what used to be easy to update manually on a white board turns out to be rather cumbersome now. The late delivery of the first batch is not yet visible on the schedule.


TAKE CARE OF PEOPLE TO BUILD TRUST

Benoît is now concluding the gemba walk. “Really interesting, thank you! I would like you to sit and think for another 5 to 10 minutes on the following two topics: first, is there something that this gemba has revealed and that you would like to dig deeper into? And, secondly, where do you need help? Please send me an email when you’re done.

With that, Benoît leaves with the management team for a thorough debrief. Guillaume, the head of the Flutter Tribe, sums it up: “Very interesting, great discussion on the misconception that piling up features in a single app would lead to a wow effect. I would also point out that we should work on the KPIs to better measure whether the app we deliver effectively matches end-user expectations. We are not at the level on problem detection and resolution. And we have project tool issues.”

Alice, the COO, concurs. “Too much time is spent on admin, not enough on problem solving. We need to help them,” she says.

“We are still far from an MVP (Minimum Viable Product),” says Marek, the CTO. “The core bid feature is still not testable at the end of August, after two months of development. They need help with the tools and training on problem solving.”

Benoît nods. “We were given a huge amount of information, which is great. I was struck by the high level of the team’s commitment and reflection. But you need to help them with their tools. The way I see it, they are both building the house and the scaffolding. We need to focus more on the code, especially because Flutter is a new technology for us,” he comments.

Benoît then shows me the email he has just received from the team, confirming the diagnosis: the team plans to challenge the customer on the release schedule, disconnecting remote live bids from the on-line features, to provide an added value to users earlier. They will also refine their measurement of user satisfaction, but request help on the tools.

The management team leaves, but Benoît and I stay put to continue our discussion.

“We are now 420 people in the group,” Benoît says. “We need to keep in mind what will commit people in a fast-growing company: in other words, what is Total Productive Maintenance for our group? We have to spend time on our basic infrastructure, systems, offices, remote work. These things can’t be taken for granted like we did in the past. This is a way to prevent failures and show respect. Not to mention keeping costs down.”


FORM A STRATEGY FROM THE GEMBA

We have grabbed a copy of the book Learning to Scale and Benoît shows me how he uses the TPS as a framework in his gemba walks. I have highlighted each stage of the discussion above with an image of the TPS house to show what he explains.

Benoît smiles. “I remember a time when I thought TPS was too far away from my digital work to be relevant. What could safety at the workstation mean to us? Or stock? We have learned a lot since then,” he says.

From the first image of the Toyota Production System, in Japanese, to a step-by-step TPS approach explicated in Learning to Scale and experimented with at Theodo, the diagram below illustrates the effort made by the group to learn, try and progressively unveil the Thinking People System – as TPS is often referred to.

I ask Benoît why he puts so much emphasis on gemba walks and genchi genbutsu. “This is sound management. If I stay in my office, the gap between my own vision of the group and reality will increase. I need to come and check regularly where we struggle and where we succeed,” he explains.

He thinks for a while and then continues: “To set a strategy for the group, I need to define where to go (point B) and how to go there (from A to B). My regular gemba walks help me understand where A is. Without a clear understanding of A, I won’t have a clear path to B.”

“Another reason,” he adds, “is the connection with the teams: during my gemba walks, I highlight the positive points, I spot talents. The time I spend on the gemba and the questions I raise show the importance of the topics we address.” Indeed, what I have witnessed has nothing to do with a traditional audit: during the gemba walk, Benoît challenges the team with questions, but never offers solutions. He tells me that those gemba walks also help the management teams: “We develop a culture of transparency, of no bullshit. It’s a good way to remain on top of the right issues, on what really matters, which can be difficult when there are many possible blind spots.”


FIGHTING OFF BIG-COMPANY DISEASE WHILE SCALING-UP

The more we grew, the more problems we had, Fabrice and I,” Benoît says. Fabrice Bernhard, you may remember from the previous article, is the co-founder of the group. “We were wondering: is this really what we dreamed of? Is this the price we pay for growth?”

Benoît confirms that Lean Thinking really gave him a method as CEO. It shows him the next step every time there is a problem, helps him and his teams not to feel guilty if any occurs, and provides a purpose and a True North. “With lean, you can do customer satisfaction and teams’ pleasure at work, or customer fulfilment and environment.”

Lean turned out to be the best way to fight off the risk of developing big-company disease. Benoît’s first steps in the job market, he recalls, convinced him that he would never want to work for a large corporation.

He and Fabrice, therefore, developed an innovative way to support the growth of Theodo. Every time they felt they had the right talent and a niche to investigate, they would create a spin-off with a pair of high-potential individuals that would act as CEO and CTO, just like they had done at Theodo. This allows them to offer a large range of products and services. Indeed, red bins on failed bids regularly reveal missing skills – whether in sales or technology – and the spin-offs are an opportunity to develop those.

Benoît deeply believes that an expert company is more visible on the market that a do-it-all group. The latest Theodo spin-off is Hokla, which uses digital expertise to track health treatments (like data mining, treatment effect tracking or chronical diseases).

Spin-offs are also a great way to retain individuals with high potential and offer them an ambitious career path as co-founders of a company. And, of course, managing small, lean, competent and autonomous organizations reduces the risk of developing a central bureaucracy.

Benoît and Fabrice may have found a lean way to growth on different market segments: much like small batches on a production line, small companies have a better chance to flow value faster to the customer.


Words: Catherine Chabiron, lean author and member of Institut Lean France