Git it done (the joy of wearing many hats)

by Giff on April 29, 2010

The good folks at Venture Hacks tweeted a piece from the WePay blog about things a non-engineer founder should know. His points tie in nicely with the struggle many business people have in taking their grand idea and shrinking it down to an actionable, minimal first version.

brickstoolsI wanted to touch on tasks for a non-coder in the super-early stages of a startup.  If that’s you, prepare to do anything and everything. Revel in the chance to wear many hats and tackle many new learning curves.  Here are some things you might work on if you are bootstrapping a company, in no particular order:

Customer Development
Get outside the building. Sure, you can organize surveys, but nothing beats a 1-on-1 conversation where you can look people in the eye. You are the sales person. You are the lean startup warrior.  Find this stuff confusing? Try reading this.

Product design
You have an idea? A startup cannot have an *idea* person who is not a *product* person.  Sketch out the desired experience flow of your application on paper.  Then make it even smaller.  Get Balsamiq and wireframe it up, because that will unearth a ton of things you have not thought about.  Have you built word of mouth and monetization drivers into the design itself?  Have you thought about an easy first user experience?

Collaborate with your technical partners, take out your scalpel again, and narrow feature set even more for an MVP (yes, it will suck, but that’s ok).  Again review with your technical partners and edit for ease of implementation both in terms of functionality and visual layout, because building a real web application is neither as fast nor as easy as most people think.  Manage ongoing requirements (although have a dev lead assign the tasks), and remember that you will be learning so much, nothing will be static.

UI Design
If you want to move fast and cheap, consider doing the first crude cut of UI design.  You will not be as good as a professional designer, but that is ok.  Study the sites you like and steal their visual ideas.  With your wireframes in hand, fire up a graphics application (I use Photoshop — just make sure you use an application that allows for layers), and design every type of screen. Review it again with your web developer for changes that could speed up development.  Then make each visual asset (i.e. a button), and write up a document that spells out everything in detail: font sizes, spacing, colors, etc.  Your job is to speed up the developer. That means removing guesswork.  To that end, organize everything clearly.  (Note: we love using Google docs now that you can use it as a shared file system)

I am not saying don’t work with a designer if you can afford one, but 1. you don’t need to be perfect in the early stages of concept validation, and 2. the critical thing is making sure you have someone who is responsive and can deal with constant change. Otherwise, take charge of this yourself.

Messaging
It is your job to figure out how to communicate your value proposition clearly and concisely.  This is harder than it sounds.  You can do this through conversations (try explaining your business to total strangers), writing and getting feedback on the marketing copy for your site, creation of a short intro video (very time consuming to do right), and even experimental ad tests on Google or Facebook.  If you want to use a blog to attract attention (for example, Mint.com), the first writer is probably you.

You should not have to distract developers to make text copy edits. Go learn about SVN (that’s what I did yesterday), get a text editor that is approved by your devs, and learn how to make changes yourself.  Do not do this casually, because you can mess stuff up, so follow directions. Learn about and respect the development process.

If you want to set up a blog, bother the devs as little as possible.  Good hosting companies (recommendations: LiquidWeb and Pair) make it easy to set up a WordPress blog. Choose a theme (I have used both Thesis and themes from WooThemes) and customize it yourself.  It will take us non-coders more time and frustration, but let your devs focus on the product. Through Internet searches and pure trial and error, you’ll get something adequate in place.

You’re not expected to transform into a coder, but as the WePay post noted, don’t be helpless either.  Whether it’s some of the basic stuff above, or learning how to access your MySQL database without mucking stuff up, minimize the requests you need to make of devs to get your own tasks done.

Smoke Tests & Landing Pages
I created our first landing page.  It was admittedly a Frankenstein monstrosity of CSS, based on my Firebug theft from landing pages I liked, but it worked (exception: I did need help in getting the javascript in place for the email submission box).  Venture Hacks also just wrote about a new service called Unbounce you might check out.  When it comes to doing any ad-based tests (such as SEM or Facebook), that’s your job as well, not to mention studying the results and explaining it to your partners.

Business Model & Metrics
Be the person that drives the lean startup process.  Lay out your assumptions and take charge of testing them.  Prioritize which key metrics are the most important things to track (the more you ask for, the more dev work is created).

Own the understanding and evolution of the business model. Understand your business ecology (customers, partners, competitors, investors, etc), expected economics, and model out a basic P&L. Study customer acquisition methods and network with other entrepreneurs to learn what is and is not working.  You will be constantly learning, so all of these things are eternal works in progress.

Networking
Don’t be an island. Talk to other startups, investors, and future partners.  It can be interesting to even talk with potential competitors.

If you are thinking about PR, it is your job to identify the journalists and bloggers you care about and try to build a relationship.

Nuts and Bolts Tasks
Your job is to insulate the programmers from corporate tasks, whether it be incorporation, setting up the bank account, tax/finance issues, dealing with the lawyers, finding office space, or writing a TOS / Privacy Policy (in bootstrap stage, DIY this by looking at established, similar websites).

Alpha & Beta Organization
You are in charge of recruiting people for your alpha and beta.  You are in charge of interviewing them.  I do not mean isolate devs from customers, but you should run the process, handle things like scheduling, and make it easy for a dev to participate but not get bogged down.

In the early stages of validation, it can be effective to wizard-of-oz your application with a manual back end.  You get to do the manual work.

You are support. You write the help FAQ. You are the community manager. You are data entry.  You are UI tester (read Steve Krug’s Rocket Surgery Made Easy).

Jack of All Trades
There are people out there who are better at each of these pieces than you: graphic/UI designers, marketing experts, sales people, accountants, etc.  That’s great, because as soon as you have some core validation to your business and some cash to work with, you can start to hire them.  But at the start, DIY most of this, and hire part-time help only where you are really falling short either by time or skill.  When hiring, remember that being good at these tasks for an established company is not the same as being good at it in the early stage of a startup, where you have to deal with constant change, uncertainty and an obsessive focus on the search for product-market fit.

I’m sure I’ve left tons out, but I am out of allotted time for this post. Back to bricklaying I go! What would you add?

  • Pingback: 6 Things a Non-Engineer Should Know Before Founding a Web Startup « WePay.com()

  • http://www.facebook.com/profile.php?id=607918 Brian Wang

    Giff, this is a great piece. I love how it neatly lays out many of the most critical non-development iteration loops that super early stage startups face. As a soon to be non-technical cofounder, I face the challenge of trying to fit some sense of structure over the overwhelming amount of work that needs to get done outside of development. I suspect that like myself, many founders are aware of what needs to get done but are sometimes unsure of where they can contribute the most in the shortest amount of time. This post really helps illuminate much of that.

  • http://giffconstable.com giffc

    glad to hear, Brian :) that's why I wrote it!

  • http://www.learnbat.com/ San Kim

    Totally. Non-technical founders need to become straight-up experts on product/UX design, to the point where they can lead, not just contribute. This is a completely separate discipline from coding, and deeply complex/nuanced one at that.

  • http://twitter.com/AndrewSkotzko Andrew Skotzko

    Love this post, Giff. Thanks for clearly articulating much of the work I'll have to do as a non-coder founder. I think this ties in nicely with much of what Sean Ellis has been posting re: organizational and business functions to get in place on the way to, and scaling out of, PMF. Can you recommend any good resources other than Balsamiq to use in helping with UI/UX?

    Thanks again,
    Andrew

  • Kelvin Jones

    Designing every screen of a web application could kill your project.

    What you want to do is design every type of screen and make sure to include every type of component that may be used.

    You developer can then use this library of components to build each screen… much quicker than having an endless number of screens.

  • http://giffconstable.com giffc

    I suspect we agree and I've slightly altered the text accordingly. Here's what I do believe, however: you should wireframe every screen unless you have an app that gets *really* dynamic. As with all things startup, advice needs to be customized to a particular context.

    Features merely as paragraphs on a page is a huge driver of scope creep because the devil is in the details. By wireframing out the user's flow, you start to see how big even your simple application is. You remember all the little admin or user tasks that got left out of the feature list. It means you can visualize, imagine and improve your user flow without having to write a line of code, saving rework. It makes it easier to figure out what to cut and/or push into backlog. And with a tool like Balsamiq, you can do this really fast.

  • http://giffconstable.com giffc

    Hi Andrew — you know, other tools like Balsamiq float across my attention but I never remember them because I'm pretty happy with B's simplicity, speed (in terms of designing — Adobe Air is a bit of a bear), and I think the hand-drawn look puts you in the right frame of mind to not get bogged down by how something looks rather than what it basically needs to do.

  • http://www.mynext.co.uk mynext

    Thank you, nice list. Have added to list of other tasks when I'm not coding. Thanks also for bringing balsamiq to my attention, first time had heard of it.

    Have a good day

  • http://giffconstable.com giffc

    Thanks for dropping by :)

  • Pingback: startupbug.com()

  • http://giffconstable.com giffc

    Off the top of my head blogs I like include http://www.slicedbreaddesign.com and http://jasonputorti.com/ and I think http://carsonified.com/blog/ is popular too
    You might follow Hiten Shah and/or Kissmetrics, and abtests on Twitter
    Smashing Magazine
    I can't remember the names of the better books, but will try to dig them up

  • http://stevecheney.posterous.com/ steve cheney

    Great post

  • http://mystartup.posterous.com Ryan

    Comprehensively excellent post. Thanks!

  • http://mystartup.posterous.com Ryan

    Comprehensively excellent post. Thanks!

  • Pingback: 6 Things a Non-Engineer Should Know Before Founding a Web Startup | Wepay()

  • Pingback: The woes of the non-technical founder « JotJournal Blog()

  • Pingback: Non Techs Starting A Tech Venture | Startup Hops()