In our Medium story Dollarize your way to better decisions, we discussed the importance of:
- Estimating hard numbers for the parameters that are important in your decisions.
- Those hard numbers need to be in common terms: dollars.
- Avoiding point estimates in favor of ranges and distributions.
- Using those distributions to get realistic estimates about the variety of possible outcomes.
That all sounds good, but how do you actually do it? To start with, you need to train your people to think in hard dollar numbers and their associated distributions. Here are three effective ways to begin that training:
Write it down
When you get an estimate on something, write it down, log it, and make it accessible to yourself in the future. Get everyone to do the same. If someone tells you that a feature is worth $250k in revenue, go back and check how many times their previous estimates actually ended up being right. It’s not that people are trying to pull a fast one, usually. It’s just that estimates need to be calibrated with actual results in order to become accurate over time. No one is born with the ability to estimate accurately.
Bet on it
Nothing sharpens the mind and the senses like winning or losing a bit of money. We’re not talking life-changing amounts here, just enough to get people to care about the result.
You: “How long is the project going to take?”
Engineer: “Should be done in two weeks.”
Y: “What are the chances it slips?”
E: “Shouldn’t slip. 10% tops.”
Y: “Ok. If you’re 90% sure you’ll make the two weeks, and 10% to slip, then I propose the following bet: if you make the deadline, I pay you $10. If it slips, you pay me $90. That would be fair bet, right? I’ll even sweeten it for you: if it slips you only have to pay me $50. If your 90% estimate is accurate, you should love to make this bet with me.”
E: “Uh. Erm. Well, maybe more like 50/50 we make the two weeks.”
Propose bets, and follow through on them. People will start actually thinking about their estimates, not just making up numbers.
The best motivator is good alignment. That means getting everyone on the same page. If your engineers don’t know the value of their work, how are they supposed to make good decisions? Err on the side of oversharing business priorities, budgeting, targets, everything. In modern tech industries, over-compartmentalization inside a company is almost always a bigger mistake than providing too much context and information. An easy way to do this is to create live dashboards that you can display on TV’s in the office. Show how much (in dollars!) is spent on features, bugs and other activities. This will get everyone comfortable with the idea.
Start with these simple ideas, and you’ve taken your first step on the path to making better decisions!