Here’s my response (slightly edited) to a question posted on the Seattle Tech Startups mailing list.  The question was: “If you were to create a startup today, what would your first 10 steps be?”

It’s not just advice, it’s what I’ve done with Resumelink in the past month.

1. Pick a working name

Don’t dwell on it, you can always change it, but you’ve got to find
something that describes what your idea is in 1 or 2 words, and can be
used as a shorthand for “my idea for a product to do X”

2. Get a domain

Claim your territory online.  This should at least be close to your
name, but don’t necessarily let domain availability determine your name,
although it can be fun to try.  You don’t need a .com — but if you
find a good one, it can be worth building a business around.  Plan to
scoop up your .com with your series A round, if your name hasn’t changed
by then.

3. Start a blog

Talk about the idea (don’t worry, no one will read your blog yet.)
This is partially about branding, you want to start building a history
of content for search engines, but mostly it forces you to think about
what you are doing.  It’s a good way to talk to yourself, and who knows,
you might get feedback.  Maybe even a partner, investor, or customer.  At
least one competitor reads my blog, and we’re working on a partnership.

4. Create a mockup & spec

Something that captures your vision (not just version 1 features).
After you’ve talked to yourself enough that you have a pretty clear
vision, try and write up something more formal.  I don’t mean formal,
just *more* formal than blog posts that occur to you as you think of
them.  This is what you’re going to show your wife, friends, and
potential partners.  You’ll probably have to rewrite it a dozen times
anyway.  Keep the spec conversational, don’t use words like “shall”.
Draw your mockups on printer paper with a sharpie, or a notepad with a
pencil.  If you want, something like Balsamiq Mockups can help, but
don’t think you need something you can pass off to a designer or
developer to implement.

5. Pick a first feature, and build a prototype/demo.

Don’t care about efficiency, beauty, scalability, security, or anything.
The idea is just to give people something to show.  Instead of filling
out a complex form, doing validation & fancy stuff on the server
side, just have a button that always shows “this is what it looks like
after you did X with a given input.”  It shouldn’t be complex.  It
should be one scenario, and shouldn’t take the user more than a minute
to complete.  Think more complex than a “contact me” form, but less
complex than Excel.  If you’re building a spreadsheet, a 10×10
table that you can edit and add formulas would be a great prototype
(just make the only formula that works “cell1+cell2+…cell9=cell10”).

6. Mention your idea to 10 people and get their feedback

Of course, you’ve probably already done this to friends and family, and
if you’re like me your wife is tolerant, but probably sick of hearing
about it in the middle of the night when she just got back from feeding
the baby.  If you’re shy like me, go ahead and count friends and family,
but pick 10 people from diverse backgrounds, not all techies, or not
only people in your model rocket club (especially if websites for model
rocket clubs is your idea.)

7. Identify 5 potential customers and 5 potential competitors

Google search keywords you would use if you were advertising.  You can
probably already identify some.  Not only will researching competitors
let you know what you are up against, but it’s something you’ll be
expected to know.  And being able to say “We’re like X because … but
we’re different from X because …” helps you clarify your market
position and is a quick way to explain it to others.  I actually
described some of my competitors on my blog and got feedback from
them.  If you can’t identify actual potential customers, by name, with
telephone numbers and email address, how are you going to sell to them?

8. Write a “pitch deck”

This is a good exercise even if you’re not looking for investors. A good
description of what should be in a pitch deck comes from the #1 Google
search result for “pitch deck”
http://startups.nuvvo.com/lesson/348-how-to-make-a-pitch-deck
Powerpoint is the original twitter, and a harder disciplinarian.  Try to
get 140 letters on a slide with 30 point font.  I dare you.

9. Get a logo and graphic design for your site

Now it’s time to start getting real.  You need to look professional.  If
you’ve got the chops, do it yourself.  But pay a few bucks and get a
website design (or at least a logo) you can be proud of.  This can be a
frustrating process, but once you are done it gives you a new level of
confidence telling people about your project — because it’s not just an
idea anymore.  There is actually something there that people can look
at, and you’ve invested real time and money.

10. Start showing what you’ve got to people and iterate.

This is where I’m at.  I’m still working on the prototype and haven’t
gotten a good web design yet, but people can register, and get a general
idea for the product.  And with a lot of manual work on the back end, I
can actually deliver the service.

Check it out at http://resumelink.org