Dan and Kate met, as usual, at Kate’s favorite coffee shop. As usual, though Dan tried to catch her arriving, he found her waving at him from a table when he walked in. They exchanged some pleasantries, and then got down to it.

“Kate,” said Dan, “I’ve got a new product in mind. I think it would bring in a few million a year at first, and probably a running rate of ten million a year indefinitely. I need some estimates of how long it would take and what it would cost to build.”

“Tell me about the product, please.”

Dan talked at length about what the market was, what the product benefits were, his best guess at the features it would need, and so on. Kate quickly picked up on the idea. She described what the organization knew about that kind of product, and identified a few areas where they didn’t have key information.

“Dan, I think I get the basic idea. I can think of a couple of your existing Product Champions who could take this one on. They’d have some learning to do, but I think we all know enough about the area that we could learn as we go.”

“Great, Kate. When can we get started and how much will it cost?”

“Do you plan to add developers, or to use an existing team? I’d recommend an existing team, because spinning up new people would be time consuming and risky. The CatBird team could be diverted to this effort in a month: we’re really down to frills on that product. We might need to do an upgrade in a year or so, but CatBird is really in good shape. That would free up Susan, who could Champion this new idea.”

“Great. How much will it cost, and when will it be done?”

“CatBird Team, as constituted, with Product Owner, is about a million dollars a year. Of course, that’s all budgeted and part of our existing expenditures. We might need some additional capital items, hardware and such, but at a guess, not much.”

Dan said, “Super. As you know, we’ve got room for reasonable capital items. How long will it take? How much total expenditure in money and time should I allocate?”

Kate said, “Intuitively, I feel sure we could have something in a year that would cover the key areas you’ve talked about. However, there are at least three big unknowns for us, as I’ve mentioned. One or all of those might be completely infeasible, or much more expensive. We’d start learning about those first, of course.”

“Yes, but how much will it cost? You’re guessing a year if all goes well. What if things don’t go well?”

Kate thought a moment. “Well, worst case, we’d find out that the product is impossible for us. Assuming that there’s little or nothing to recover, we’d lose whatever investment we had made.”

“How much would that be? How much will it cost to decide whether we can’t do it at all?”

Kate said, “At a random guess, a month for the whole team, for each of the top three risk items. Three months total in risk assessment. Over that interval, we might get components to use later, but we might not see an integrated product at all. More likely, we could focus part of the team on a thin product, as we always do, with another part working on the hard technical bits. There’s not much advantage to throwing all ten people at the hard bits.”

Dan looked worried. “So it’s possible that we could burn $300,000 and come out with nothing?”

“Yes. If we start with the hardest obstacle first, and work on the others a bit, we might bring that down. But I don’t see any way to be sure we’re not blocked in under two to three months’ work.”

Kate went on. “Much more likely, though, is that we would realize we’re blocked much sooner. I think the chance of discovering a complete washout is 5% in the first month, 3% in the second, less than 2% in the third. We’re pretty good at chasing down what we don’t know.”

“So you think there’s a 10% chance we can’t do it at all?”

Kate leaned back to think. Perhaps coincidentally, the coffee shop chatter stopped for a moment, as if there had been a disturbance in the Force. “Intuitively, yes. I was just trying to express that we’re most likely to discover problems early, and that as time goes on, less likely to hit a complete obstacle.”

It was Dan’s turn to think. The coffee shop didn’t seem as interested in his posture as it had been in Kate’s. Finally he spoke:

“Kate, I trust you, you know that. I’ve got two concerns. First, based on what we’ve talked about, we could completely lose up to three hundred thousand dollars. That’s a lot. Second, we just have an intuitive notion that we could get a reasonably viable product for about a million. I’m concerned about what happens if we can’t.”

Kate said, “So what do you need?”

Dan said, “I think I need a more firm estimate of the risks, how much we could lose entirely. And I need an estimate of the features we need, how long they’ll take, and what they’ll cost.”

Kate said, “OK, I have two questions. First, how much money can we spend to answer those questions? We’ve never built a product like this, so while you have a vague idea of the features, no one else does. Which should remind us to be sure you dedicate time to bring us up to speed on our vision. We should start on that right away, as soon as team and Champion are available.

“Second, how much time are you willing to delay inception while we gain more certainty?

“Third (I lied about the two), can we use the CatBird team and my recommended Product Champion to do the work needed to get these estimates?”

Dan said, “I’d rather have CatBird continue to provide new features. Can’t you estimate this some other way?”

Kate said, “Well, I just pulled these numbers out of my sweater. To get better numbers it seems to me we need to apply real product expertise and real technical expertise. That’s got to come from somewhere. CatBird is the most available team.”

“What about in their spare time?”

Kate tilted her head. “I don’t know that phrase. Explain it to me?”

“Well,” said Dan, “they could do some stories from CatBird and some from this new thing.”

Kate said, “So you’re thinking slow down CatBird, so as to get better answers somehow faster than we could get from actually working on the new problem?”

Dan said, “Have I mentioned lately that I hate you? But really, can’t we get better numbers more cheaply?”

Kate said, “Back to my questions. How much of your time can we have? How much money can we spend to get better numbers? How long are you willing to delay dropping the flag to get more information? And I’ve decided if you ask me to do this, I’m dedicating CatBird one way or another.”

Dan thought. “What can you do for $100,000?”

“One month of waiting, free of charge, while CatBird untangles. One month of CatBird, doing their best to figure out if we’re technically blocked. In parallel, one month of Product Champion, working with CatBird, to flesh out features. Two month delay, $100,000 completely at risk. At the end we’ll have rough cuts at technical feasibility for our top concerns, and a bitty little product increment. We can meet then and tell you what we’ve learned, and you can decide to stop or to go ahead.”

Dan said, “I need a better deal. I need estimates with no expenditures and less delay.”

Kate sat back and said, “With whom outside our organization should I set up an appointment for you to pitch this idea to? This is the best we can do.”

Dan said, “But I need you to do this. I need these numbers. I need to be sure.”

Kate said, “In two elapsed months, with $100,000 invested, non-refundable, we can give you more certainty.”

Dan said, “I need it in one month for $50,000.”

Kate said, “Stop CatBird now, six weeks, $150,000. Lose a few features in CatBird, and tick off Susan for cutting the project out from under her. I can probably resolve that by making her Champion for the new thing. Things in CatBird will be left hanging and everyone will feel jerked around. That will slow them down. It’s six weeks instead of a month, because my proposed month before inception gives us time to settle the waves.”

Dan said, “That’s a worse offer than your previous one.”

Kate gave her most evil smile, which wasn’t all that evil. “True. Want to play again? Your best bet, in my opinion, is to wind up CatBird in a month, take another month burning $100,000, and re-assess then.”

Dan said, “OK, let’s do that. Deal?”

Kate said, “Deal. Just one more thing.”

“What’s that?” Dan asked.

Kate said, “We just agreed on exactly the same thing I initially suggested.”

Dan displayed confusion.

Kate said, “Think about it. We’ll wait a month, start burning at a rate of $100,000, and show you what you’ve got at the end of the first month. By then, we’ll have reduced about half of the technical risks, built a mini-product, figured out a rough backlog. We’ll probably have a couple more uncertainties that may need resolving in the next couple of months, but of course we’ll have a better sense of those. There might be more uncertainties, there might not.

“You’ve always had the ability to stop a project whenever you want, and to receive the full benefit of the investment. In early months, that’s learning and a mini-product. Later on, it’s more and more product as well as more and more learning.”

Dan thought. “You minx! It is the same, isn’t it? We begin investing to resolve uncertainty. We learn by building the thing. We can stop early if it’s not working out and carry on as long as it seems to be paying off.”

Kate smiled. “Yep. That’s what we do here. Welcome to the new Oak River Software.”