Saturday, 1 August 2015

"What can I get for..?"

I've been following some discussions on twitter and read in various blog posts about shifting to budgets (, I've even written some about it as well: It's not really a shift, because I think budgets always exists in one way or another, but the idea seems to be to move to questions like "What can I get for $X?" or "What is possible for $X?" or even "How much are you willing to spend/invest?", "How much is this worth for you?" (or similar).

These are seemingly great questions (approaches) and it may work really well in some situations. But I sense some problems with it as well, let me explain.

In everyday life, this is probably how we buy things. If I'm interested in e.g. a new TV, then "What TV can I get for $X?" is a common, sensible question. And also a good thing to let the TV salesman know. But buying software isn't really the same thing as buying a TV or a car or whatever. As the famous quote goes "Your job isn't to get the requirements right - your job is to change the world" (Jeff Patton). In other words, we're not selling cars - we're selling "dreams" (sounds epic, right?)
I think, perhaps, it's a better analogy to compare buying software to buying art or (even better?) a pet.
If I want a dog, the main question in my head isn't "What kind of a dog can I get for $X?" or "What is a dog worth to me?". And if a dog breeder told me "How much are you willing to spend on a dog?" that would - to me - come off as a pretty weird question to ask. I have "a dream" of owning a Labrador Retriever (just an example, and I think you can compare with art here as well), not just settling with a dog I can afford (although that might be something someone actually would do). I think the same goes for software in many cases. (I wrote an article in a Swedish computer magazine about the analogy of software and pets: "Så får du it att räcka vacker tass" (Swedish)).

In my experience, the "value part" or "what is this worth?" is quite often not really known (or very hard to predict) or perhaps not even really interesting. Of course, getting to know these things might be worthwhile, and sometimes (often?) it's actually really good to find out/explore these figures. But again, the most interesting question when buying pet/art/software often isn't "What is this worth (in $) to me?". Thus, the first question isn't "What are you willing to spend?" (the car salesman approach). The first question to ask is "What is your 'dream'?".

Then, of course, the inevitable question will come: "What will this cost me?" and whatever we respond, the response to that most likely(?) will be "I can't spend that much!". But what if someone doesn't? Wouldn't you rather have someone having their "dream" fulfilled (or as close as we can get) over getting to know, upfront, what they are willing to spend on something? (I'm stretching the word "dream" here, I know, but I think you get it). And to be honest, I believe the question: "What are you willing to spend?" (or similar) most likely will make you come off as a "car salesman". Not a good start (if you start with that question).

Of course, in business, the difference is that we want to make money. Art/pets won't (not usually, at least) let you do that - and it's (usually) not the goal either. But again, that question is quite often really hard to know; "How will this make me (a lot of) money?" And here I believe that moving to validating assumptions (instead) is a good approach. Like: "How can I grow a sustainable business out of this?" (read more in The Lean Startup. Although there are things to consider as well: Lean Start-Up, and How It Almost Killed Our Company). This is another topic.

Another quite common thing is that some things "must" be done, for various reasons (e.g. regulatory), and then the question: "How much can I get for $X?" isn't relevant either.

Also worth considering is that being explicit with your budget might not be that great (in a "low trust environment"?):


  1. There are three probabilistic elements oif all projects. If you fix the budget, the other two are still in play. Since they are probabilistic, estimates are needed to determine there values and the probability there is sufficient budget.

  2. Thanks for the post. Looking at the pinned Tweets by Peter (and your comment), like you say, this only applies in a low trust environment. That's not an environment in which I want to live (ideally) because it causes all sorts of dysfunctions.