Twenty-eight years from the studio to the product roadmap

Mandate

Turn an app built around devoted audiobook listeners into a service that could grow with lighter, older and Spanish-speaking listeners without royalty costs outgrowing subscription revenue.

Outcomes

• Grew paid subscriptions from 0.98 million to 1.58 million, including 240,000 in Spain after the 2022 launch.
• Replaced the single unlimited plan with hour-based plans in March 2023; gross margin rose from 31% to 38% by 2024.
• Rebuilt the home screen into labelled sponsored, editorial and personalised rows; independent publishers' share of impressions recovered from 8% to 16%.
• Launched Mode lecture in November 2024; monthly churn among subscribers over sixty-five fell from 2.6% to 1.8%.
• Launched the free podcast tier in March 2025: 2.3 million monthly listeners and 61,000 paid conversions by June 2026.

Decisions in context

Lisandre recruited me when it had just under a million subscriptions, one unlimited plan and an app whose every screen assumed a listener who finished a book a week. Those listeners had built the company, and by 2022 they were also consuming its margin. My first year went on research and on the Spanish launch, which taught us how much our recommendations depended on French listening history. The four decisions that followed were the hour-based plans, the separation of the home screen into rows with named owners, Mode lecture for older listeners and the free podcast tier. Each went to the executive committee with a dissenting memo attached, and two of them went to the board. I own plan structure but not prices, which I propose with the finance director, and I own the recommendation strategy but not the model itself, which belongs to the recommendation science team inside my organisation. I am not a director of the company.

Reason for the move

Current appointment; completing the family plan and the Spanish catalogue work.

Mandate

Rebuild the customer app and website so that customers could understand their bills and change their contracts without calling, and manage the streaming bundles sold with tariffs.

Outcomes

• Redesigned the app around the bill and data balance; bill-related calls fell 27% by 2020, saving about €5.6m a year.
• Grew monthly app users from 1.1 million to 1.6 million; the app-store rating rose from 2.6 to 4.3.
• Moved 85% of contract changes to self-service by 2020, from about 30% in 2016.
• Replaced automatically started streaming bundles, 41% of which had never been activated, with customer activation.

Decisions in context

Merovia was my first large company and my first contact with telecoms economics, where a small change in a self-service journey moves millions of euros of care cost. I joined to run the customer app, the website and the content partnerships, which bundled third-party streaming services into tariffs. The app's problem was not technical. It had been designed as a sales channel and was used as a statement of account. Turning it round meant giving up in-app sales, which put me at odds with the marketing director for most of a year, and telling customers when a cheaper tariff suited them, which the chief commercial officer allowed as a pilot and the board's customer and commercial committee later made permanent. On bundles I learnt that an activated subscription and a counted one are different things. In November 2020 Lisandre approached me while I was leading Merovia's negotiation with it; I disclosed the approach the same day and handed over the file.

Reason for the move

Recruited by Lisandre Audio to lead product; I withdrew from Merovia's bundle negotiation with Lisandre as soon as the approach was made.

Mandate

Build paid digital subscriptions for the group's newspapers alongside print, without losing the loyal print readership.

Outcomes

• Grew digital subscriptions from 9,000 to 71,000 between 2012 and 2016.
• Linked 48,000 print subscribers to digital accounts through a telephone activation line.
• Replaced a ten-article meter with marked subscriber articles, raising monthly conversion from 0.4% to 1.1% of regular visitors.
• Launched a daily local news audio briefing, built on Ondaline's technology, with 85,000 weekly listeners by 2016.

Decisions in context

I arrived at Hérisart through the acquisition of Ondaline and stayed to build something I had never built: a paid subscription. The group's newspapers were losing print revenue every year, and their most loyal readers were also their oldest, many of them subscribers for decades who had no intention of creating an online account. Our first digital product asked them to do exactly that. The telephone activation line, staffed by former subscription clerks who knew the readers by name, did more for digital growth than any feature we built. I also replaced the metered paywall with subscriber articles chosen by editors, after noticing that the meter mainly taught casual visitors to clear their cookies. During these years I completed my executive master's part-time, which gave me the finance vocabulary I had lacked. I left when Merovia offered the chance to work on a product used by more than a million people every month.

Reason for the move

Recruited by Merovia Mobile to lead consumer digital products at much larger scale.

Mandate

Turn a prototype for replaying local radio programmes into an app that stations would pay to join and listeners would return to.

Outcomes

• Grew monthly listeners from about 15,000 to 420,000 by 2011.
• Signed 61 local and community radio stations as paying partners.
• Built the app's first recommendation feature and a weekly digest for occasional listeners, which doubled their return rate.
• Prepared the product and data for the sale to Groupe Hérisart.

Decisions in context

Two former Radio Clairval colleagues founded Ondaline to give local stations what national broadcasters were starting to offer: a way to hear a programme after it had aired. I joined as the first product manager, mainly because I had run the network's podcasts and knew what stations would and would not do. I learnt product management by doing it, from specifications and café usability tests to pricing station partnerships. My main mistake was building for our most devoted listeners. The next-episode feature delighted people who already opened the app daily and did nothing for the occasional listener, who was most of our audience. A weekly email digest, written by a human editor and tuned by listening history, fixed more of that than the algorithm did. When Groupe Hérisart bought the company, I chose to go with the product rather than look elsewhere.

Reason for the move

Ondaline was acquired by Groupe Hérisart in January 2012, and I moved into the group's digital team.

Mandate

Take over a neglected website and give the network's programmes a life after broadcast.

Outcomes

• Launched the network's first podcasts in 2005, a daily local news bulletin and the documentary slot.
• Reached about 160,000 podcast downloads a month by 2007.
• Published an online archive of 3,400 programmes with searchable summaries.
• Persuaded six stations to record local council meetings for on-demand listening.

Decisions in context

The web role came to me because nobody else at the network wanted it, and because I had been asking awkward questions about how listeners could hear programmes they had missed. For two years I edited the website, which mostly listed schedules. From 2005 I ran it, and the first decision I made was to publish our documentaries and a daily news bulletin as podcasts, against the advice of the sales team, who feared that on-demand listening would reduce live audiences. It did not, as far as we could measure. The archive we built became a resource for schools and local historians, and the council-meeting recordings were among our most downloaded items, which taught me that usefulness beats novelty. I left when the founders of Ondaline offered me the chance to do this full time for more than one network.

Reason for the move

Joined former colleagues at Ondaline as its first product manager.

Mandate

Report local news for the Lyon station and produce the network's weekly half-hour documentary.

Outcomes

• Produced 96 half-hour documentaries between 1999 and 2003.
• Won a regional radio documentary prize in 2002 for a series on Lyon's last silk weavers.
• Answered more than 300 requests for a replay of that series by posting cassette copies.

Decisions in context

My first job was the one I had trained for: reporting council meetings, floods and factory closures for the Lyon station and producing a weekly documentary for the whole network. Radio is unforgiving editing. A listener cannot turn back a page, so every sentence has to earn its place the first time it is heard. I learnt to interview people who had never spoken into a microphone and to cut a forty-minute conversation into four minutes without betraying it. The silk-weavers series in 2001 was the most important thing I made, less for the prize it won than for the letters it produced. They showed me an audience that wanted to choose when to listen and had no means to. When the station asked me to take on the website in 2003, I accepted because of those letters.

Reason for the move

Asked to take over the network's website as web editor.