A couple of posts back I mentioned Estimation and my desire to poke a stick at the hornet's nest that estimation can be. Estimation is always a controversial topic. It's often at the heart of serious conflicts within organisations. There are a huge number of estimation methods and techniques but nothing seems to prevent these issues from coming up. Before I poke a stick into the hornet's next (well, not so much poke as take a full bodied swing with run up and follow through), I'll spend a little while looking at why we estimate in the first place.

Any time we have two parties involved in something there is estimation happening. Right back from prehistoric times -

Ogg - I estimate that this stone axe is worth the same as that reed bag filled with nuts so let's trade.

Ugg - I estimate that this reed bag filled with nuts is worth at least 2 axes so let's not.

Ogg - Seriously? You're taking food away from my family... these axes are the finest workmanship. Maybe one axe plus a flint scraper but that’s the best I can do.

And so on.

Every time we conduct a transaction with someone else, we estimate the value of what we are getting and compare it to the value of what we are giving in return. Fundamentally, we want to make sure that we are getting a good deal. We don't like getting ripped off.

 The one difference we have now is that often, one party in the transaction knows exactly the value of what they have because it's measured in Money. The other party is offering a good or service, so there has to be an estimation of whether the value of that good or service matches the value of the money being offered. This is where we start to run into problems.

Often the benefits in a transaction are intangible - a good night out, a tasty beverage, an improved user experience. The difficulty comes when we try to convert that intangible value into a dollar value. People are strange sometimes. Because money is tangible, people put a far higher value on it that it really deserves. Psychologists have done studies on this. If you offer people a $20 note or a bottle of wine that you tell them is worth $20, almost everyone will take the money. Even if you tell them that the bottle is worth $30, most will take the money. It's not until the bottle of wine is well over $50 that people will start to prefer the wine over the money. That's because money has a definite value but the value of wine is intangible.

So when we estimate, what we are trying to do is put a dollar value on something. This is complicated by people over-valuing money and under-valuing intangibles. "But wait" you say, "we don't give dollar estimates, we give estimates in time". Well, yes, but the person you give the estimate to immediately does a conversion - "4 weeks of developer time at $150/hour..." and will arrive at a dollar value. The upshot of this is that it usually falls on the party offering their intangible part of the transaction to somehow "prove" that their estimate of value is correct.

This is also complicated by a power imbalance in the transaction. Usually the person offering the money has more power than the person offering the good or service. If we take our prehistoric transaction as an example, this power imbalance often looks like -

Ugg - No. This basket is full of the best nuts of the season. Definitely worth two axes.

Ogg - You are starting to annoy me. And I have an axe. So hand over the basket...

Of course, when an estimate turns out to be incorrect - it takes 4 weeks instead of 3, the wine wasn't very good or the nuts were slightly mouldy, one party feels ripped off. They didn't get the value they were promised, so next time they need to do business they will drive a much harder bargain. They will start to demand "more accurate" estimates to avoid being ripped off again.

This is where we get into the very serious business of estimation. The whole thing basically boils down to putting a value on the intangible half of a transaction and coming up with some way to prove that that valuation is correct (and prevent the person in power hitting you with their axe). Because people are generally very bad at estimation, over the years those on the money side of the transaction have been (in their view) ripped off so often, that they drive a very hard bargain indeed and demand a huge level of accuracy in estimates to prevent it happening again.

Over the years we have come up with a huge variety of methods for making our estimates "more accurate", which is code for "we can prove it will take us this long and cost that much". Unfortunately, most estimation methods don't increase accuracy. What they increase is precision which is a different thing. And we will look at that difference next time...