The Program Wasn't Broken. It Was Filed Wrong.
A failing program rarely needs what everyone reaches for first.
At Teradata I inherited an executive briefing program that ran like a coin flip. Whether a session landed depended on which account team showed up and which experts wandered in from Product and Engineering. The instinct in a spot like that is to rebuild. Better tooling, tighter prep, a sharper team, more discipline on execution.
That instinct was wrong. The program wasn't broken. It was in the wrong place.
Two moves
It lived inside Product Engineering, an org whose job was building the product, not running the highest-stakes conversations the company had with its customers. So I moved it to Marketing and rebuilt the function around briefings, customer stories, and references. Same program. Same name. It became the company's top source of marketing-sourced pipeline, over a billion dollars influenced, with an NPS of 97 against a world-class bar of 80.
Later I moved it again, to Sales, where it sits today. Each move was the same question asked one more time: whose outcome does this function serve?
Product Engineering
Builds the product. Briefings were a side duty.
A coin flipMarketing
Briefings, customer stories and references run as one function.
Proved the pipelineSales
The room where executives decide is where deals are won.
Where it sits todayRebuilding versus recategorizing
Here is the distinction that matters. Rebuilding is a management move. You execute harder inside the frame you were handed. Recategorizing is a leadership move. You see that the frame itself is wrong, and you change it.
Most leaders claim the first and leave the second on the table, because moving a function admits someone had it filed wrong, and that is socially expensive. The redesign is the easy part. Saying it out loud is where most people stop.
The same error is all over AI
I see it in AI initiatives right now. The stalled one usually isn't a model problem. It is owned by the wrong team and measured against the wrong outcome. A customer-facing assistant owned by a team that is measured on uptime. A sales tool owned by a lab that is measured on demos given.
Before anyone rebuilds, run the cheaper test. Ask whose outcome the thing serves, and whether that is who owns it.
Most stalled initiatives aren't broken. They're filed wrong.
If you run customer programs or executive engagement and you are working through this inside your own team, I'd like to compare notes.