The Headcount Metric
The unintended consequences of valuing managers based on how many reports they've had.
I once interviewed at a company that was looking to replace their head of engineering. The biggest issue with the previous one was that he was solely focused on adding to his headcount. Increasing headcount was more important to him than growing the business or making it more attractive for a fundraising round. The company suffered for it.
What was one of the first questions I was asked? What was the largest headcount I had managed.
To quote the economists out there: incentives matter. When one of the first questions for any management position is "What was the largest headcount you managed?" and it is widely known that that is one of the first questions, what do you think most people in management positions will try to do?
They are incentivized to add to their headcount to make their resume look better. Sometimes it is knowingly at the cost of making a sub-optimal decision for the business. Sometimes it is done unknowingly by finding a justification for the increased headcount. This doesn't make them bad people. It makes them human. Humans respond to their incentives.
One thing I always found interesting before I entered management was the prevalence of a per head budget in addition to an overall budget for the engineering team. It made no sense to me at the time. If you have a really good engineering candidate, who can do things no one else on the team can and can likely help the rest of the team grow, why not increase the offer by 20%-30%? If you have 15-20 engineers on the team, one really good engineer with the expertise to raise everyone's productivity is well worth 2 engineers who are just extra capacity. Spending more on fewer heads makes perfect sense.
It is incredibly rare to have the opportunity to put hard numbers on that concept, but I had one in my career. I'm changing the numbers for anonymity.
There was one time I inherited a group of outsourced engineers. 4 of them were assigned on a front end application that was being developed by the time I arrived. It took them 3 months and was a disaster for the most part. The dependency tree was a complete mess. The code structure was incomprehensible. Nothing would have been transferable to a mobile application so one would have to be built/duplicated from scratch. These engineers cost a total of $12k a month for a project total of $36k on something that was barely releaseable.
I hired a great frontend dev who cost us $15k a month. A single engineer that costs more than 4 other engineers sounds extreme... until you look at the results. This engineer onboarded, analyzed the existing codebase (realizing this was one of the few instances where throwing it away made sense), and rebuilt the entire application in 6 weeks. The project total was $22.5k. Not only would this have saved us close to 40% of the original cost, but we would have gotten a product out much sooner. There's opportunity cost in 6 weeks. We also had a better code base which would have made it easier to work on future iterations of the product. It also was built in a way that eased porting it to a mobile app so future dev costs were lowered even further.
The only downside was the idea of spending more on a single developer than 4 developers.
Interviewing for management positions made me realize how big of a downside that is for a lot of people. The need to fill out the headcount resume line item is why we don't just have engineering budgets, we have per head budgets as well. It's why there is more focus on the cost of an engineer rather than the capabilities of that engineer. It's likely why the salaries of engineers aren't more reflective of their contributions.
I've never cared about inflating my headcount. That isn't because I'm a better person, I'm not. I simply have different incentives. I have always wanted to be an entrepreneur. The only metric that really matters when it comes to engineering for an entrepreneur is the return on investment. Every other metric is supporting or serving as an indicator for ROI. Headcount is not a metric that helps with ROI.
Ultimately, all businesses want an ROI. The problem is when we confuse what is and what isn't a useful metric for helping us get to a better ROI. Headcount makes sense at first and only starts to look shakey when analyzing the work engineers manage to do. It's a lot more effort, but shifting conversations from headcount to ROI will eventually get us better engineering leaders and better run engineering orgs.