Custer, South Dakota, September 2024
Subscribe now
People ask me how Broadcom grow so quickly, from an edgy startup to an $8B company. But I can only tell you the role I played:
Building out a brand-new way of using data for decisions.
Many of those people remember that era like this: every manager – top to bottom – could see the effect of their decisions on the company’s overall financial plan. And they all could update the plan with new news about the business when they discovered it, giving the CFO instant visibility to the P&L impact of every decision, analyzable by any angle and any time horizon.
Imagine the power of decisions in that setting.
I’d be lying if I said that my team’s innovations started with a great vision and technical skills. Truthfully, the data solutions we used in those early days – before AI or big data – were more like survival tactics. We were soldiers fighting a new kind of war, inventing new weapons to keep from getting killed. What really drove my team were the high expectations of an army of PhDs. Our leaders wanted visibility into the project work of 4,000 engineers, stated in financially accurate dollars, looking two years into the future, all tied to the sales pipeline. And they demanded all of this immediately.
Risky Business
I still remember reviewing this project with a senior VP. It seemed like he was thinking, “My team is inventing the internet. You think what I’m asking you to do is hard?”
A bold move was our only option.
I canceled a contract for the software tool we were using (which the engineers hated) and put all bets on my team to deliver the solution. We built an input form directly in our data warehouse, a counterintuitive approach in the IT world. It worked; the engineers embraced our new tool to plan their time, and the product teams used the data to make great decisions about their roadmaps.
That solution wasn’t just a technical innovation. Managing plan data directly in the analytics system created seamless continuity between all the historical transaction data in the business systems and the planning data in the minds of the executives. Decision makers got real-time feedback on the overall effect of their inputs, and historical data (which only changed daily) matched up perfectly, giving context to all their inputs.
That removed a mountain of friction between Broadcom’s business processes and its thought processes.
Things really took off from there. Soon, hundreds of product managers began forecasting their own revenue. Engineers projected manufacturing yields. Finance analysts tracked data for new product proposals. All of these inputs were combined into a real-time P&L forecast. Complete accountability was normal – everyone knew it – and that led to very accurate projections.
The Root of Innovation
What made it possible for the company to move forward with this new “frictionless” approach?
It was the willingness to start over.
If the high expectations of senior leaders hadn’t challenged me in the first place, I probably wouldn’t have recognized the friction our first solution caused. The lesson I learned from these engineers speaks to the core of all innovation: when you’re willing to throw out your own solution and start over, you really don’t lose much. Instead, you gain full awareness of what doesn’t work from that first, painful experience.
Maybe that’s where you are right now. If you’ve got a data solution that isn’t cutting it, incremental fixes won’t solve the problem. A house with a crumbling foundation doesn’t get stronger by painting the walls. When it comes to a data challenge, your willingness to start over usually leads to a better solution.
Stop defending bad data solutions. You don’t need a new BI tool; you need a new data strategy.
Thanks for reading! You can find more ideas for your data by visiting my Substack site.