Invup documentation

It must have mattered. Look how bad it was. Not only were the pages just ugly, they were peppered with database errors and CSS blowups

. It doesn’t prove design doesn’t matter, but it does suggest design may not be the deciding factor.

Can you see at what point in time we changed design?  No?  Must not have made a difference. Let’s look at time-on-site:

Nothing. But this is all superficial — what Really Matters is the Conversion Rate: are more or fewer people signing up each week:

Hmm. Looks like everything objective is saying “it doesn’t matter.” But as much as I respect and follow Lean Startup theory, objective measurements aren’t the only things that matter. Those customer emails matter too.

The other day we landed a large customer who said they could tell from our website that among our competitors we’re more mature and ready to handle a bigger customer like them. I can tell you — objectively — that we’re among the youngest of our competitors, and although I have a list of reasons why “we’re better,” the truth is that particular customer would probably be served just fine by several of those competitors. Was it the design that gave us that edge? Could be. Didn’t hurt, anyway.

Still, the more I look at the importance of design in the startups in my little career, the less it seems to matter.

Modern Lean Startup theory blares out from the red-tiled rooftops of Stanford: Seek the Data and Ye Shall Find! First the bounce-rate. If our website design was repulsive — literally — the bounce rate should now diminish. Here’s the data:

An even more extreme example comes from my second company ITWatchDogs. I displayed its old homepage at the magnificent Webstock design conference in Wellington earlier this year; the crowd whooped at our violent assault on the visual arts, complete with calliope menubar colors, two broken images tag above the fold, and a layout model that could be seen as a “grid” only after consuming a pillowcase of mushrooms:

But you’re anticipating the punch-line — ITWatchDogs grew every month, made millions of dollars, stole business from competitors with billion-dollar market caps (and professional-looking websites), and had a successful exit.

Of course it’s only fair to also point out some of the many instructive counter-examples: Hipmunk is the same thing as Orbitz or Travelocity — the only difference is amazing design, not just because it looks good but because it’s so useable. In the words of Joel Spoksly — the design “affords usability.” (P.S. Early Hipmunk team member Alexis Ohanian is so cool and smart and rich and funny and successful and good-looking that really he doesn’t deserve to be alive. (P.P.S. Hey flamers, for God’s sake it’s a joke! Don’t you realize I’m just sore from losing the Pecha Kucha competition to him?)) I always use and recommend Amy Hoy’s time zone tool only because it’s just nice to use and look at. (P.S. she also authors a terrific blog aimed at the solo entrepreneur.) Many people credit Mint’s smash success with their terrific design. Considering how many features were broken for how long, it’s hard to argue. 37signals documented — with data — how design changes results directly in more credit-card-swiping customers. It doesn’t get more “business value” than that. So where does that leave us in the “matters / doesn’t matter” question of design?

I think you can go either way, but you must decide whether or not you’re going to value design as core to your startup’s identity, and then act consistently. Here’s what I mean.

It’s clear from the outset that design is the only competitive advantage Hipmunk has over its competition. Specifically, by making the flight-search problem pleasurable and useable instead of feeling like you’re navigating pivot tables from Excel ’98.

They don’t have better data, better branding, better name, better SEO, or more money. Just better design, and not just easily-copyable incremental improvement, but a quantum leap better.

When design is that fundamental to the business — how it acquires and retains customers, garners attention and referrals, and distinguishes itself in the market — obviously design can be the most important thing.

Conversely, at ITWatchDogs the company’s internal and external culture was that we’re low-cost, friendly, approachable, regular guys, who understand exactly what you worry about, exactly what your budget is, and we nail it. The site might have looked bad, but our message couldn’t have been clearer.

But it is useful to decide where you come down on the question of design in your startup, because if it’s important you’d better work on that right now and develop a consistent culture of valuing design through-and-through, and if it’s not important you’d better decide what is important and nail those things all the harder, because you’ll be competing with people who are using superior design to cover up their lack of competency in those same areas.

4. Optional: Determine which outlet gets the scoop In today’s media landscape, a story covered 15 minutes ago by another blog can already be considered old news. Therefore depending on the type of news and the market you’re in it might be wise to give one specific media channel the scoop.

8. Distribute your press release In a previous step you set up draft emails for all your PR contacts, now is the time to send them.

6. Write draft emails In the previous step you have probably only sent out the news to a handful of contacts on your contact list.

7. Make the announcement When the day finally comes, you can publish your press release (or have it published automatically if you’re using the schedule option).

5. Send personalized emails to journalists, under embargo If you decided not to give the scoop to a certain blog you can still let journalists know about your announcement before you actually make it public.

9. Thank the people that covered your story Congratulations! Now that your press release is out in the open you should see some coverage happening. Be sure to enjoy all the free publicity your company receives and thank the people that cover your story via a personal email so you can start building a relationship with them for your future press releases.

The other option, and probably most common, is to have someone or several folks dedicated to “product management.”  That probably makes sense in larger organizations or as companies grow,

However, I personally think that this is very dangerous for an early company, and I’d go so far as to say it’s a major red flag.  Here’s an example of what you want to avoid

it’s a good sign and highly preferable if the founders focus and own the product process.  Ideally they’re also able to “go deep” and be involved in the entire product management process, writing the specs for engineering teams, managing the priorities, etc…

I think the founders of a startup need to be very cognizant and controlling of the product management process, at least until you become a larger organization, and avoid the temptation to hire “someone dedicated to product.”  In my experience, the more people/layers you add to the process and the further the founders step away from the product management the process, the worse the outcome.

The trouble with my last company was that our experience in software development came from large companies, where your job is just to ship code.

So we thought that a launch was just about writing the code.  And, we did that in 6 months.  But, what we didn't account for was that in a startup, you don't have a ready large group of users just waiting to use your product. So, your launch time must also include a cycle of user experience and marketing.

So, if you do everything in series in a drawn out way like we did: build, market, measure, it's a cycle that can turn months into years.  Eric Ries suggests that shortening an iterative loop and going through such a loop multiple times quickly is the key to success.  I would take that a step further and suggest not only cutting activities to shorten that loop, but to do as much of this loop in parallel.

Our workflow looks like this: get your Unbounce or LaunchRock page up from Day 1 and start marketing before you have a product.  You can gauge interest and get signups from the very beginning until you're done with the first iteration of the product.  Start getting the Craigslist posts out there on Day 1 to get feedback and potential customers immediately.  Once you have enough of an idea of what to build, start mocking up your idea.  Get those mocks back out to potential customers to make sure you're on the right track.  Iterate as much as possible on paper before building, because it's much faster to re-draw than to re-code.  "Delete features" on your paper prototypes as well, reducing what you need to actually build in code.  Try to code as little as possible to shrink that build time to about 1-2 weeks.  By the time you're done building your first prototype, you've already acquired users from doing marketing in parallel.  This puts you in a position to start measuring usage and gauging interest immediately before iterating through that loop again.

Build, market, measure should happen as much as possible in parallel to reduce your launch time and keep your money, morale, and support up.