Things to do before you[r project] die[s]

Feb 17, 2017 00:10 · 1430 words · 7 minutes read IoT Kickstarter

It’s no secret that all projects die, eventually, and that most projects die before they have had a chance to have any impact. This is true of personal projects, of open-source projects, of startup projects, of hardware projects, and more.

A few days ago, the IoT project on which I was working did just that. It was funded, it was open-source, it was open-hardware, and it still died before having the opportunity to release anything useful. In our case, we got pretty far, didn’t run out of funds, and were a little polish away from putting our first alpha-testers in front of a pretty advanced prototype before a higher-level strategy pivot killed a number of projects at once.

Ah, well, risks of the trade.

Still, working on it, we managed to learn or come up with a few tricks to increase the chances of surviving at least long enough to release a product. I’ll try to summarize and explain some of these tricks here. Hopefully, if you have a project, regardless of funding and medium, you may find some of this summary useful.

Death by Nobody-will-use-your-project

tl;dr

There are many good ways to test an idea way before the prototype is solid enough to show it up. Use them.

You have an idea. It’s certainly a great idea. Your friends and your family probably love it. A few geeks around you are enthusiastic, it can change the world!

At this stage, you are eager to start coding/designing/burning circuits/… And that’s great. By all means, if you’re enjoying yourself, please carry on. However, before you start committing to the project anything you might regret – leaving your day job, spending your money or your investors’ – you should make sure that your project has a chance to gain users and traction.

If you come from a technological background, you may believe that you need to put your project in front of users to get feedback, which means that you need a pretty advanced prototype. If you do this, you considerably increase your chances of your project dying on you.

So time to put your Survivor Hat.

There are many ways to get user feedback while the project is in its infancy. Doing this used to be long, expensive and – just as important for an indie project – boring. These days, there are a few services that do it for you and actually bring money to your project in case of success. Yes, I am talking of Kickstarter, Indiegogo & co. If you believed that these crowdfunding platforms were about gathering money, welcome to the worst-kept secret of the 2010s: they are actually marketing platforms that will let you find out if people are willing to spend money on your project. Which, in many cases, is a pretty good measure of whether you should greenlight your own project. Just be sure to be honest regarding the current state of your project and how the money will be used. If you lie, even with the best of intentions, your funders will eventually find out and they will be legitimately angry.

An alternative, depending on what you’re doing, is to open a website and start taking pre-orders (morally dubious), start letting interested parties subscribe to an announcement mailing-list (better), or get creative. I have even seen weirder variants in which non-technical founders started a fairly sophisticated online service before they even hired a single developer – by hiring underpaid students to actually do the work of the machine, at least long enough to determine whether users would actually use them (neat trick, morally… weird). I have also seen at least one startup-turned-internet-giant who implemented only a subset of the promised features and measured how many users clicked on buttons that were not implemented yet (not a trick I like, but it worked for them).

Note that all these techniques help you determine your potential end users. Unless your end users are developers, that’s entirely orthogonal to getting people to take part in the development of your project.

This list is clearly not exhaustive. Now that you have your Survivor Hat on, you should be able to come up alternatives if you need to.

Regardless of the technique you use, spending a few days of work to measuring your potential audience can save you from disastrous choices for your career (or savings, or startup, …). As a bonus, putting together some communication will also serve to put your own thoughts in order, decide of your priorities, get feedback from your future users. Also, if you need money, it can serve to fund you (crowdfunding) and/or to convince investors.

Death by Tests-need-users-need-tests

tl;dr

There are many good ways to test algorithms, UX, accessibility, … before the prototype is solid enough to show it up. Use them.

At this stage, I’m assuming that you have identified your user population and the conditions in which your project will be used. You will, of course, need to adapt – you will need different testers for an OCR-for-the-blind or for a superbowl-adblocker. For narrative purposes, though, I’ll pretend that your testers somehow lurk in your hallway.

So, your users are interested, your project is going full speed ahead. Great. But, like it or not, full speed ahead is often still too slow with regard to testing. Why? Because just as code doesn’t work until it has been {unit, integration}-tested, a product doesn’t work until it has been user-tested. Again, if you are thinking as a developer, you will be tempted to wait until you have an advanced prototype. But, if you are thinking as a developer, you also know that you don’t wait until the project is nearly complete to start adding {unit, integration} tests. This is doubly true if your project involves hardware – just building the hardware in the correct form factor can take months that you typically do not have.

So put on your Survivor Hat and find ways to run user tests on a prototype that doesn’t exist yet.

UX can be tested on mockups. If it’s a GUI, a text interface or a braille interface, draw it using any mockup tool, or even on paper, grab some people in the hallway, see if they can guess how to go from “I want to do this” to “I’m clicking the right button”. If it’s a sound/voice UI, hide your mockup, grab some more people in the hallway, and roleplay the speech recognition/text-to-speech engine, etc. Even wearable IoT devices can be mocked up using wire, velcro, a smartphone and some bluetooth accessories.

Similarly, you can user-test algorithms. Add a mockup UX on top of it – again, it can be as simple as paper – grab some people in the hallway and run the algorithm. If the algorithm is simple enough, you can again roleplay it. If it is more complex, turn it into a command-line/REPL toplevel and run it behind the scenes.

Let’s go further. Are you building an IoT device? If you are lucky, you can quickly rig together a prototype running your code on top of a smartphone, smart tv, alexa or any existing device that your future users already own – and if no smart device has all the capabilities you need for testing your code, maybe a Bluetooth thingy can provide the missing features. Or perhaps you can prototype your software application as a browser add-on. Or detach a single feature and turn it into an App. Suddenly, instead of having access to a handful of potential alpha-testers, you can put your code in front of potential millions.

Again, regardless of the technique you use, spending a few weeks on mockups/early prototypes will let you perform user test in parallel with the unit and integration tests that you already run continuously. Much as unit/integration tests, user tests will let you fix issues before they kill you. Hey, sometimes, a prototype can even become a product.

Now what?

Well, hopefully, this blog post gave you a few ideas. Now, put on your Survivor Hat and start thinking about ways to make your project survive until it has a chance to have an impact!

As a side-note, while I came up with the “Survivor Hat” as a narrative to explain some of the techniques above to fellow developers and while we came up with some of the ideas above mostly independently during Project Lighthouse, this blog entry is largely about some of the “Lean Startup” concepts. So, if you are interested you should find it easy to read more on the general idea.