I’ve been asked some variation of this question a lot recently:
“I have a simple product idea that I want to try out. I want a MVP version, nothing fancy, just enough to prove the value and raise investment (or get a green light from my boss for the rest of it). How much will it cost you to build it?”
The implied question behind this is, “A MVP should be cheap, right?” Before any discussion has taken place, and almost irrespective of the scope of the product, non-technical people frequently have the mistaken idea that their vision can be brought to life in rough form within a few weeks and for under $10,000, maybe $20,000 tops.
Here’s why that’s crazy and what you should expect to spend to realize your vision, assuming that neither you nor, in the case of a startup, anybody else on your founding team can code a single line of it.
It will take (a lot) longer to build than you think
You’ve got a “design document” or something similar. It looks so simple on paper! And you’ve heard the tales of legendary engineers who bang out algorithms while sipping Red Bull on 18-hour coding stints. Wasn’t Napster, like, built in one weekend or something?!
Ask a junior engineer how long it will take to build something based on a verbal description or written document, and he’ll give you an optimistic answer that you like. Ask a really great senior engineer the same question, and she’ll give you an answer that makes you blush. She knows what the junior engineer doesn’t, which is that it’ll take more than twice as long as she’d think based on the document. And that your document will be ripped up halfway through.
Even the simplest web or mobile application will take 4 to 6 weeks to reach a demonstrable state. And I don’t mean something that looks a lot like what you envisioned. I mean a crude, squint-hard-and-you-might-see-it version that has one or two the features implemented in a rudimentary way. A good analogy is touring a house that has only the framing in place. You can see where the walls and roof will go – and now you’re suddenly realizing that the master bathroom is way too small! – but that’s about it.
If you have a really good engineer, he’s going to recommend at this point that you spend a week or two working on automation to make each iteration of your vision faster to build and deploy. This should be non-negotiable, but it’s frequently skipped in order to finish the “critical” work, which is perhaps the worst false economy in software development.
Now you can iterate. You’ll build in the rest of the features, go back and forth on your design, tear up the design document, have multi-hour discussions about the goals and which things can be cut or which things are suddenly must-haves. You’ll spend 2 to 3 times as long in this phase as you spent in developing the first demo, so for the purposes of our example, let’s say 8 to 12 more weeks.
So you’re looking at 3 months minimum to have a product that’s good enough to launch as a MVP. If your product has any complexity to it at all – anything with AI, big data, asynchronous processing, significant third-party integrations, etc. – then you should expect 2x, 3x, or 4x that, and know that the risk of never getting to the end is pretty high.
Good software engineers are expensive
I’ve written previously about this. In short, the demand for senior software engineers is very high right now. Even an average senior engineer in Austin, Texas will command a salary north of $120,000/year. Good ones have a floor of $140,000/year. Before benefits, cash bonuses, ISO’s, PTO, and other ancillary perks.
Multiply the salary by 1.5 (at least) to account for the all-in hiring costs – employment taxes, benefits, equipment, software licenses, possibly office space – and you have a ballpark of $180,000/year minimum to employ one of these people. If you then figure it will take 3-6 months to build your MVP, you could estimate spending “only” $45,000 to $90,000.
Again, this assumes you’re just going for an average senior engineer. You can spend more up front for a great engineer and end up with a better product, more quickly, and with a lower long-term TCO. This is a simple expense now vs. expense later tradeoff.
A lot of companies want to make the tradeoff in the other direction: save money now, spend more later. This lines up well with a startup’s financial realities, for example. But it’s a siren’s call in most cases. Your options for saving money now are either finding a junior engineer or going off-shore. I’ve never seen either of these options work out well for a MVP, not just in terms of the quality of the finished product (which will be poor) but also in terms of the cost to get there (because the iteration time is so long). I’m not saying it can’t be done – I just haven’t seen it.
If you’re a startup, the only other way to save money now is to hire a good senior engineer for a low salary by compensating with “lots” of equity. This is entirely possible to do, because most engineers have a terrible mental map of the expected value of that equity. You can find someone who will take 0.5-1.0% of your 2-person startup that has zero revenue/customers/users, no serious investment, and no provable product-market fit and trade that off against $20,000-$40,000/year of salary. They’ve read too many rags-to-riches stories to say no. Bad for them, fantastic for you!
“Too many words! Just give me the executive summary.”
No matter how simple you think your product is, count on $50,000 at the low end for a MVP. And please understand that this is a ridiculously optimistic estimate.
If you want to be sort of safe, assume you’ll need at least $250,000.
Chances are good that your number falls somewhere in there.