Kickstarter and ship dates

It seems I mostly use this blog to rant about unreasonable expectations, and I’m going to continue in this fine tradition, so here goes:

It really pisses me off when people complain about ship dates on Kickstarter. Some of the complaints might include:

• You missed a shipping estimate, so you’re a crook and I deserve my money back!
• You changed your shipping estimate part way through! Is outrage!
• You’re refusing to give us a ship date? That’s bullshit — you obviously know one, so why not tell us?

When you join a Kickstater project, it’s in the beginning of its development. The creators probably have a reasonable idea that what they want to do is actually possible (although they might not be 100% certain!), but they certainly won’t know every difficulty they’ll encounter along the way. At that stage, nobody can know for sure when it’ll ship. The creators aren’t being arses and withholding information — that information just doesn’t exist. That’s why if they give a date at all, it’ll change. It’s not bad management if it changes, that’s the reality of product development.

Consider big tech companies. Most of them wait to announce their products to their customers until they’re very far through development. Apple don’t announce until they’ve done their initial production line run and have them sat in a warehouse. That way, they can be pretty confident they’ll hit their target ship date. Microsoft announced the Surface ages ago but only gave the ship date very recently. It’s all to avoid missing dates, since the date can’t be known until late in the development cycle.

Kickstarter cannot function that way. The money is needed earlier, to be able to commission assembly lines, order components, and other such things. But as I mentioned above, the date can’t be known until after all those things are done. If backers think they have a right to know a shipping date from the beginning, it’s a catch-22: something has to give. Either they get given an inaccurate date (and complain when it changes), or they get given no date (and complain about that fact).

Why would the dates change? There’s two reasons, and one applies more to Kickstarter. Firstly, I’d wager that lots of people on Kickstarter aren’t very familiar with the realities of commissioning production, or of just how long shipping thousands of boxes will take, or of various possible delays in the process. That’s not their fault, they just haven’t done it before. It can (and should) be mitigated by doing research before they begin, but not everything can be anticipated that way.

The second reason is just that estimating how long things will take is really difficult. I’m a software engineer, and I hate it when our client manager asks me to estimate how long it’ll take me to write a certain feature. I want to give him a huge range, like “between 3 days and 3 weeks”, because if it’s as simple as I hope it will be then it’ll be done in three days, but if it requires rearchitecting part of the core of the product to make it work, it’ll be the longer one. Why don’t I know already? Because I haven’t written it yet — it’s only by actually doing the thing that you know which end of the range was correct.

But clients hate that kind of estimate. Remember, it’s just for one feature — if you multiply it out for all the features, you get an app that’ll be written in somewhere between 3 and 24 months. The only accurate date we could give would be “Less than 2 years”. You can’t tell them that. So we average them, and hope that if some things take longer then others might be quicker than expected and it’ll all balance out. That’s dangerous, because we don’t know for sure that we can do everything in time, but it’s the way the client software industry works: clients want a time estimate, so we try to make one. If occasionally we miss the estimate, then discover that the client turned our “Hopefully we’ll be feature complete by November” into “Lets spend our entire marketing budget on putting stickers on a print run of books that ship November”, that’s why we have a full time client manager, to deal with the fallout and let the techs keep working on making the thing.

So for big companies announcing things to consumers, they don’t mention the existence of the thing until they get to the point where they’re sure of the ship dates. For small companies doing work for clients, they give a best estimate of a date while knowing things could go wrong and take longer. Nobody gives the actual estimated date range, because the response is “That’s a huge range! You must know something more specific!”

What should Kickstarter projects do?

Sadly, there isn’t an easy answer. Pebble, the project that prompted this post, originally gave an estimate of September. They’re in a good position compared to most projects, since they’ve actually made a smart watch before — lots of people are coming in completely fresh. Anyhow, they gave that estimate before the project got funded, then they got $10,000k instead of $100k. That is, one hundred times the amount of demand they’d required as a minimum. At that point, they were out of their experience zone, since manufacturing doesn’t scale linearly.

After that, they steadfastly refused to list another shipping date, responding with some variant of “it’ll be done when it’s ready”. I think this is the right thing to do. They’re giving us progress updates (we just had photos of X-rays of various manufactured samples to check for defects), so we know they’re actually working on it. But saying they don’t know the actual ship date is probably true. Why would they know it, when they’ve never done production on this scale before? It’s not like there’s a book that has a big table of project requirements to ship dates. So hats off to them for not just pulling a number out of their arse to (temporarily) shut up the whiners.

The best thing I think a future project could do would be in the original pitch, to say something like “If we get under 1000 backers, we will aim for a December ship date. Be aware that this date will probably change as we encounter unexpected situations during the project.” Note the deliberate absence of another estimate for if they go over that number of backers. There will still be whiners if they do change the date, but at least firstly they’ll have made it clear that the date is likely to change, and secondly they’ll have a complete get-out if the project is more popular than they expect.