A traditional company that had grown by buying others
More than 11,000 employees, over 90 subsidiaries, each brand with its own way of working. Departments were siloed with goals that did not line up. A top-down project mindset made it hard to react to the market or to a customer. There were no standard metrics and no shared access to the ones that existed.
With the Chief Growth Officer I helped restructure product management itself. The Chief Growth Officer who hired me described the result in his recommendation as “structuring the entire Product Management function leading to a way more mature setup around defined KPIs (OKRs), clear roles and structured communication.”
Defined roles and accountabilities
Every role got a written scope and an accountability.
Ambiguity about who decides is the most expensive thing in a product organization and the cheapest to fix. It is also the thing nobody fixes, because writing down who owns what feels like administration, right up until two people both think they own a decision and neither makes it.
The Product Trio
Inside each team, product manager, designer and engineer own value, usability and feasibility risk together.
Three named risks, three named owners, one decision. The value of the model is less the collaboration and more the naming: when a launch goes badly it is clear which risk was not covered, and that is a conversation a team can have without it becoming about people.
OKRs in the open
Tracked transparently in Atlassian Atlas, so a team in one brand could see what a team in another was trying to achieve and why.
Goals that live in a deck are goals nobody can act on. Goals a neighbouring team can read are goals that stop two brands solving the same problem twice, which in a group of 90 subsidiaries happens constantly and invisibly.
A shared shape for the work
The Double Diamond as the common structure for a project, Teresa Torres’ continuous discovery habits inside the teams, and Scrum supported by agile coaches.
The point is less which framework and more that everybody used the same one. A group assembled by acquisition has as many processes as it has acquisitions, and the cost is not that any of them is wrong. It is that nobody can move between them.
Structured communication
A working model for where decisions get announced, where questions get asked, and where the record lives.
In a group of 90 subsidiaries the alternative is each brand inventing its own, and none of them reaching each other. This is the least glamorous item on the list and the one that turns the other four into how the group works.
Decisions
Write down who owns what, before the first new ritual. Planned: begin with the visible things, Scrum and a discovery cadence, because ceremonies stand up fast and show progress inside a quarter. Argued: a ceremony is a container. Run one in a group where two people each believe a call is theirs and the meeting becomes the place the call is avoided. The dull paperwork has to come first or the visible work has nothing to hold. Happened: scopes and accountabilities on paper, then the Product Trio, and the ceremonies afterwards. By then most of them picked themselves.
Put the goals where a neighbouring brand can read them. Planned: report goals upward inside each brand, which is how a group assembled from acquisitions normally handles it. Argued: upward reporting answers to the people who already know. The expensive problem in a federation is two brands spending a year on the same thing without either one finding out, and nothing fixes that except making a goal legible to a stranger. Happened: Atlassian Atlas, open across brands. It is also what made a shared shape for the work worth agreeing on, because a goal you can read is a goal you can recognize as yours.
The honest caveat
The shift to Scrum sped up delivery in most teams and not all of them, and maturity was still growing when FTI Group filed for insolvency in June 2024, for reasons that had nothing to do with the product organization.
Culture change that is not measured has to be re-argued every quarter. This had better stories than numbers, and I would put a figure on decision quality in the first quarter if I did it again.