- project, product, company. Before getting too excited about your
latest idea, ask yourself if it's a project, a product or a company.
A project is some useful and innovative tool or service, but unlike a
product, it's unclear if anybody will enough pay for it to justify its
manufacture (and delivery, i.e. through channels) -- much less whether
this is still true in the presence of substitutes and knockoffs. Products
are projects that are sellable-- they have financing, designs that
incorporate feedback from prospects or customers ("people who can pay
enough"), reasonable quality controls and processes, legal representation,
A company is a product with headroom: infrastructure to grow, a
fleshed-out management team, the ability to create multiple products, etc.
Since companies can be tiny and have one product, the real distinction
between "product" and "company" is whether the financial returns of the
product justifies the corporate structure necessary to make it successful.
A classic problem is starting a company that doesn't attract enough
capital or talent to give it life beyond the founder's personal
investments in time and money, thus starving the company of longevity, the
ability to weather storms, the plethora of different skills to execute on
the vision, etc.
- the gorilla game. Another thing to consider is whether your project
is suited for a startup. For example, database management systems (DBMSs)
make for lousy startups, because (these days) they're a feature of a product--
not a whole product that a customer will buy. It would be a tough road to
sell jet engines in the after-market: buyers expect them to be deeply
integrated and supported by the jet manufacturers (Boeing and Airbus).
Another version of the problem comes in scale-- big companies own
distribution channels and partnerships and branding that can impede your
success. That's why it was so difficult to compete with Microsoft in
desktop applications, and why it's so painful to get "shelf space" in
larger retail chains, including groceries.
Being on the wrong end of the gorilla game isn't fatal, and can be
rewarding. For winner-take-all markets, the game plan is to partner with,
or get acquired by a large company which will complete your product.
To do this, you need to demonstrate both grass-roots success as well as
a fit with the larger companies products.
Finally, history is made by the people who violate this and other
"rules"-- from TiVO ("DVRs are a feature of your cable company"), to
Google ("who needs another search engine?"). The key is having a truly
breath-taking product that's so compelling that customers and partners
ignore what you lack. It's crucial that your product can be "purchased"
separately-- it would be a tough road for a startup to sell jet engines in
You'll know when you have it, because people "in the know" will tell you:
they've tried other solutions and choose your product and use it all day
- be real. A lot of the time, you're playing poker-- for you hold'em
players, we're talking big bets on the river card, when you can't push
anybody off the table. Like any good poker player knows, you don't want
your bluffs called very often. Technically, it's fine once in a while
because it "keeps'em guessing"-- but that's a very advanced strategy I
wouldn't recommend for first-timers. Once your reputation is sullied,
it's hard to get back.
For example, if you claim that your technology does "ten times" better,
you'd better be ready to prove that... in someone else's lab. If you say
you've got N million users, make sure you've double-checked your log
analysis-- or better yet, have it proven through Google Analytics, Nielsen
Another tip: defend yours and your company's reputation and do so honorably.
When competitors or executive knock you down, take the high road and use it
as an opportunity to sell yourself to them and others. As one example, I
was once chastised by a middle manager for flubbing an executive presentation--
where I hadn't been the one presenting! Many things I write off, but not
this: I've given >100 public and executive preso's and press interviews, and
I always practice beforehand and always read the papers and watch the videos
afterwards. Moreover, it's a key part of my career and it would be dangerous
for this mistaken identity to spread. Incidentally, since he was my manager's
manager at the time, and this was the straw that broke the camel's back, I
immediately put in a request for a transfer-- who knows what other bum raps
he's giving me? post-script written 18 months later: that manager was later
stripped of managerial responsibility and fired a year later.
- what to worry about. Carefully pick your investors: even if you
control the board, crappy investors make your life miserable and distract
you from success. Down the road, they can gang up on you to make a bad
decision. Finally, they can say the wrong things to other people
e.g. employees, customers, competitors and other investors. Even the best
investors say and do stupid things all the time-- focus on whether they'll
do brilliant things and are well connected.
Carefully pick your employees-- in most cases, they are the company
(not code or patents). Eventually, the company can manage turnover, but
even then it's painful.
Pick your lawyers carefully-- I've heard horror stories of mediocre
lawyers letting startups get into trouble. Go with a firm (and
individuals at the firm) who have lots of startup experience and lots of
current startup clients in your field. Healthcare looks nothing
like enterprise software. You will need many types of lawyers:
transaction lawyers for negotiating and closing deals; corporate lawyers
for investments, compliance and day-to-day business; IP lawyers for filing
patents. Good lawyers should "get it" and also give you sound advice that
you can double-check. Get referrals, hold out for good ones, check
references, and make sure you have good bi-directional communication.
Good lawyers are in demand, be prepared to pitch them. $400 per hour is
small potatoes compared to defending a lawsuit or signing a bad deal.
Carefully pick your customers-- they need to have a real demand for your
product; be educated about alternatives; can speak on your behalf to
investors, employees and other customers; be technical. Ideally,
customers are well known inside and outside their industry, willing to pay
lots of money and represent an example for which there are many others.
About all of these-- always be willing to walk away if you get a 'funny
feeling'. When the opportunity smells like it'll "take the company to the
next level," make sure you have good lawyers help craft the deal.
- outsource HR. HR scales up but does not "scale down," and
therefore it's a great candidate for outsourcing. I prefer TriNet which we used at Inktomi,
Addamark and others. They really "get" the startup game. Email me, and
I'll get you a referral to someone there. Another company is AdminiStaff but (circa 2004) they
seem less friendly to startups.
- recruit for skills, not titles. For the first couple of years,
you're going to be gated on the availability of skills in the company, and
people will have to "wear many hats." This has many implications. First,
you want to recruit people with many skills and who are comfortable
wearing many hats. A CEO who only knows how to sell isn't much help-- the
person also needs to manage. A tech person who can only "do software" is
a classic disaster-- she also needs to know how to raise money. Second,
you want people who aggressively (not just "willingly") handoff roles to
specialists over time, esp. "outside their organization". Power politics
is destructive. Third, you want to be careful not to "hire your friends".
This is tough: only your friends will be "crazy enough" to join your
startup, but too many of them and your company will lack the breadth of
experience to survive paradigm shifts. For example, if your programmers
all prefer to work on Windows, but the market is demanding Linux, you'll
be screwed and not even know it until you try to port the code (insert
horror story here).
One trick I first saw at Perforce: don't give your executives VP titles
until the company is ready. Hire them as Directors, and when they push
back, explain that the VP title is being reserved for when the company is
big enough to justify it-- and it's their job to make that title come to
life. This avoids title inflation, with a 100 person company having SVPs
and EVPs and CxOs. There are ways to assuage fears about "hired over,"
but the best is to avoid having VPs at all. Finally, the best executives
will be there for company success, not for the title.
- order of hiring. When you first get going, the order in
which you hire people is a critical success factor. First, you don't have
the resources (capital, wallclock time, and managerial time/energy) to
hire (and train/integrate) people you don't absolutely need. In most
cases, capital constraints make this self-evident-- $1m doesn't go very
far. But even if you're well-funded, hiring people in the wrong order
will leave you trying to keep high-powered people busy, create unnecessary
corporate goals (aka "politics"), etc.
Without further ado, my preferred order of hiring is:
1) prototype product engineer. You need someone who can bang out a
prototype, suitable for raising money. This is typically a full-time
job-- often the future CTO-- but I've seen it be contracted out to
part-timers. Many successful companies have been founded by
businessfolks who recruited someone who did this.
2) early stage product manager. Next, you need someone (or a team) who
knows how to design a prototype. Almost invariably, this is one
of the founders.
3) general business manager with startup experience. Once you have
a prototype in the works, you need a business-focused person who can
pitch it to others and guide it into being a commercial venture. Core
skills and experience should include: startup financing, startup
fundraising, marketing, negotiation, startup legal issues, startup HR
issues, etc. These skills are needed even to support product
4) legal. You need this early on, and it should be outsourced. If you
have a good idea, there are terrific lawyers at classy lawfirms, who
specialize in representing startups-- they take a percent of your
company's stock, in return for deferring their (exorbitant) legal bills
until you get financing.
5) business development. As soon as the prototype is fleshed-out enough
to sell, you need a busdev person to get some customers, if only to get
feedback about the product. Revenue also helps to close financing.
You don't want a "bag carrying" salesperson-- they're too
expensive (in cash), tend to be market-specific and they rarely have
the skills to "invent" sellable products from your prototypes.
Less-early-stage hires include: quality assurance (QA) and engineering
management, marketing, office manager, financial management (controller or
It is worth noting that hiring-order is sloppy stuff: invariably, you have
times when you're under-staffed and/or high-powered people aren't busy enough.
- the turnover paradox. IMHO, turnover should be absolutely
minimized. First, it makes for a crappy culture-- a revolving door
reminds people of the impermanence of a startup and it makes people feel
personally at risk. Second, recruiting takes a lot of time;
turnover doubles that. Third, it means that your intellectual property
and/or business ideas/lessons are walking out the door, possibly to
That said, the turnover paradox is that double-digit percent turnover (per
year) is healthy. Between the stress of hiring quickly enough to succeed,
limits on who'll be crazy enough to join, and constant shifts in strategy,
it's impossible to know who will work out. Therefore, my advice is (a)
avoid "firing" people-- lay off the position(s). First, this reduces the
risk of lawsuits; second, it reduces the stigma to them; and third, it
places the blame on your own head where it belongs-- a "bad hire" is
always the fault of the manager, never the employee. Even if the
person misrepresented him/herself, it's your job to call references and
double-check this. You should be thinking that 100% of the responsibility
is your own, and self-assess after each turnover. (b) Try to make it
mutual and not stigmatic. Some people will "blow up" on you, but
generally it's possible to wave goodbye. For example, I hate
cliff-vesting because it shifts the responsibility in the wrong direction,
and doesn't save enough stock to be worthwhile. Having former employees
hold a little stock (and be proud of having worked at your company) is a
great way to reduce the risk of them selling your ideas (aka consulting)
to competitors. (c) remember that high-powered talent hates to be
under-utilized. Though people often don't see it this way at first, it's
better for the person in the long term to get out of a bad situation. (d)
turnover is critical for attracting, retaining and motivating top talent
and reducing politics. When underperformers are coddled, the
overachievers start whispering in the halls and whining in private. This
is no fun for anyone.
My philosophy on startup performance reviews: (a) they should be tied to
events, not the calendar (although in sales, the quarter is the event).
(b) they should be 360-degree both to avoid protection/favoritism, as well
as to capture team effects, (c) they should be simple/numeric and not
longhand-- both to reduce conflicts, but also to keep people from spending
lots of time thinking about writing (many people hate to write). Also,
90% of review content people already know-- so it's an abhorrent waste of
time, for something people hate to do.
Instead, I prefer something like this (modified from Cohera): the basic
review is a 10-scale form across each of 10 department-specific axes, each
person eval'ing N=4 other people around the company plus themselves. The
CEO gives a number of points you could "spend" on your evals-- e.g. if the
team were executing like crazy, we'd get 300+ points to spend across the 4
evals. Among people who score really high or low vs. their role, more
detailed reviews were done. Likewise, more work was done when
self-assessment was wildly out of sync with outside assessments.
Low-scorers could request that additional assessments be done. This way,
we created a review funnel, allowing people to focus their energies on
work, instead of writing performance reviews. The points can be doled out
per department, ie. if sales is sucking wind, it gets fewer points...
- separation of role. It makes sense to hire a lawyer to represent
you in court-- even if you're a trained lawyer. Likewise, there are
numerous times when you can derive benefits from having two people work
together, even when one of them has the skills to handle both roles.
"Good cop, bad cop" negotiation is a classic example. Another example is
in demonstrating to investors that you have empowered people representing
conflicting roles, e.g. product development versus sales.
- early VC fundraising. Fundraising is a form of sales. There are
a million books on raising VC money, so I'll skip to the things they don't
say. About VCs: (1) most VCs are terrible at "picking winners" and even
the reasonably good ones are only good in narrow niches. VCs don't
"develop ideas"-- they evaluate whether a proposal is good or not, and
don't have the patience or time to "fix" proposals that are slightly off.
Fortunately, good VCs won't stigmatize you-- I once raised a round from a
VC who was referred by another VC-- even though the original VC turned us
down. When the funding announcement went out, the original VC sent a very
nice congratulations email. (2) It pays to learn the dynamics inside VC
firms. For example, VC firms typically require two senior partners to
drive a deal through. Junior partners don't count, for example. Thus,
once you woo a senior partner, expect to need a second one before the
partner meeting. In fact, the referral story in (1) came because we found
a VC who "liked our deal" but couldn't interest his partners, so he
referred us out.
- angel fundraising. Again, I'll skip the basics e.g. raise only
from qualified investors, etc. (1) Angels invest for all sorts of
reasons. For example, tech-savvy investors often invest for the prestige
and for the networking benefits; marketing-savvy investors often invest
for the pride/prestige of "picking winners" (yes, you can "win" yet return
little to the original investors); wall street people and bankers often
invest for the excitement. (2) There's a lot of paperwork in raising
money, and we're talking about qualified investors ($1m net worth) so you
should set some reasonable floor, at least $10K. Going the other way, if
it takes 2 hours to raise $30K, you're still beating the dollars-per-hour
efficiency of raising from VC firms (in a series A). (3) Don't expect
much from friends-of-friends -- it's hard enough to get non-trivial sums
of money from people who've known you for years, much less strangers. (4)
Because of #3, you'll quickly exhaust the low-hanging fruit, leaving only
longshots and long-cycle (6mon+) opportunities. One solution is to
recruit key executives on pure-stock and deferred-comp packages, then
leverage their rolodexes to raise additional money. In one venture I
know, the visionary founder raised half the money, the cofounder raised a
third and one of the early executives raised the rest. The ultimate VC
financing came through a lead by that early executive. (5) Avoid putting
in your own money. A certain amount is healthy, but it's very bad for the
company to be funded by insiders-- you want outsiders rooting for your
success. Outside investors also provide a more objective view on your
company, and act as a measure of success. Having many "smart money"
outside investors looks great; having a CEO self-fund the company buys you
nothing. Finally, outside investors help with leads-- sales, fundraising
[Aug'09] I'm often asked about raising angel financing, there are
several useful tactics. First off, treat this like a sales funnel, and
expect an order of magnitude higher hitrate among direct friends and
former colleagues of the founders, than among friends-of-friends and
friends-of-non-founders. Second, a helpful tactic is to not "raise money"
but instead "ask for feedback"-- this takes the pressure off potential
investors, who can ask you if they can invest. Third, a
particularly efficient route is to hold "feedback sessions" among "friends
of the company"-- the goal is to fill a room at the private home of a
successful entrepreneur, with 15-20 people who are "heavy hitters" either
in tech, your startup's domain area, or business. Think famous named
people and/or famous companies they work(ed) for. The idea is that person
X will attend to meet and impress person Y, i.e. this is a networking
event. Doing this ensures high quality feedback: everybody's on best
behavior and using the feedback as a way to network. Again, you're
obviously raising money, but it's all done softly-- interested investors
will approach you afterwards one-on-one, comforted that they're not alone
in supporting your company. Fourth, a "technical advisory board" and/or
"customer advisory board" is a great way to show thought leadership, and
also blunt worries about competition, i.e. if famous person X is
associating herself with your company, then you must be a leading innovator.
- recruiting. [Apr'04] Recruiting is the single biggest determinant
for success, with the people you hire literally being the DNA of the
company. The challenge of recruiting is often under-appreciated by people
who've never been responsible for it. First, you need to "sell" people on
joining your venture, which isn't easy. Remember the old adage "good help
is hard to fine"-- that's because smart, reliable, hard-working,
no-nonsense people are never without work, so you have to lure them away.
Here's some basic tips: (1) strategize: list reasons why people would join
your venture instead of others, including reasons they wouldn't. Then,
use this to source candidates. For example, when talking with recruiters,
let them know these things. When considering how to source candidates,
emphasize channels that tend to fit your criteria. (2) treat hiring like
a project, including realistic schedules, budgets, risk analysis,
contingency plans, etc. This avoids a lot of heartache when candidates
ask for a lot more money, can't join when you need them, etc. (3) use
pipeline management-- an ordered set of steps, with tracking of how
candidates are getting through the process. You can optimize the pipeline
by front-loading the steps that are most likely to filter-out candidates,
and which are the least time-consuming. A proper pipeline also lets you
optimize each step separately. For example, my first contact with a
candidate is designed to be cheap-- usually an email with some basic
information. I call candidates who express interest-- sadly, I haven't
found a way to avoid these bazillion calls, since I don't like to
pre-screen too much.
(4) respect legal and ethical guidelines. Long before lawsuits or
violence, hurt feelings are common in hiring. Since life is long, you may
run into candidates in the future, where they remember your rejection.
The world is a small place, so be careful of your reputation in the hiring
process. Ditto this advice in day-to-day management after hiring.
If you're recruiting for someone else, remember that you've built a
relationship with the candidate, and should be available if problems arise.
I like to check in from time to time.
(5) tips for checking references. Of course, you check references, right?
good. And you don't hand this to HR, right? good. They ask dumb
questions like "how long did you work with X? what are X's strengths and
weaknesses?" These questions have obvious "correct" answers, which
neither avoids fraudelent references, nor elicits the issues that would
lead to not hiring the person, much less provide information on how best
to manage the person after they're on board-- which is the real value of
reference checking IMHO. And HR won't impress the reference-giver, who
will often talk with the candidate afterwards about their experience.
Instead, you can chat up the reference-giver and impress them that (a) you
know the candidate and can articulate why you're excited about working
with the person, (b) your company and department are a great place to work
and that (c) you'd be an awesome boss. Oh, and this is also an
opportunity to source candidates for other positions. Some good reference
- what project(s) did you and X work on? This detects fraudulent
references, and also tells you how sharp the reference giver is, which
IMHO is a good gauge of whether you should trust this reference-giver.
Make sure the reference describes project(s) in enough detail.
- what was X's most shining moment? This helps gather stories, which
you can then use to butter-up your management if you need to ask for
deal-terms (e.g. money) that are outside your parameters. This question
also butters-up the reference-giver for more difficult questions, such as...
- can you describe a situation in which X failed to deliver on his/her
goals? Then followup with "how long did it take X to realize s/he
wouldn't succeed?" and "how did X communicate this?" and "how did
management react?" This detects "kicked-dog" syndrome, and lets you know
what sorts of sensitivities the candidate might have, and how they might
react in a politically-charged situation.
- can you describe an corporate culture or environment that X would NOT
be successful in? You may have to push a little and pidgeonhole the
reference who's intent on blowing sunshine. It helps to ask this before
you describe your company, of course.
(6) interview tips. Obviously, respect is the key here. Don't show up
late, take a phonecall or eat a burger during an interview. And treat
every minute as precious and not recount stories from childhood. Don'y
ask stupid HR questions like "describe your strengths and weaknesses"?
when you could instead ask "what do you like about this position and why
do you want it?". Don't ask a easy or trivial technical questions when
you can instead ask deep, hard technical questions that detect junior
vs. senior talent and detect incompetence. This is best phrased without
acronyms or jargon. For example, my favorite question at Addamark for
engineers was "managing big data-- say several terabytes-- is a big pain
in the neck. Can you describe some of the challenges?" Junior people
will talk about technical details, senior people will talk about
fundamentals. Incompetent people will get those details wrong.
There are a slew of miscellaneous tips I've found-- YMMV. (1) I try to
schedule interviews for after lunch, say 2pm. This avoids wasting all day
in a first interview, avoids taking candidates out to lunch, which
empirically seems to hurt the hit rate (sorry, I have no explanation), and
also uses the staff when they're least productive and most reasonable,
i.e. while digesting burritos. (2) ask a person the range of cash
compensation they want in the first phonecall, apologize for the
awkwardness, and do it before saying anything about you want to offer.
This avoids interviewing people you can't afford. Also, some people will
name low-ball numbers which could either be a good or bad sign. If the
number is in range, I tell the candidate up front, which puts people at
ease and lets them focus on the job itself, rather than compensation.
(3) stock compensation is messy because candidates either under-value or
over-value the stuff, which creates deal-closing and post-deal management
problems. (4) In an hour-long interview, I look for the candidate to tell
me something (relevant to the job) that I didn't already know.
- corporate structure and captable. [Apr'04] Consider the corner
restaurant that runs out of money before they figure out their pricing,
fix quality problems -- and build up their stock of "regulars." What
happened was that they were under-capitalized, and it happens to high tech
companies all the time. There's an old silicon valley saying: "always
raise more money than you need." Riffing on the adage "a banker is
someone who gives you an umbrella when it's sunny and takes it away when
it's raining," I often say "a venture capitalist is someone who sells you
an umbrella for $1 when it's sunny, but $1000 when it's raining."
You can also raise too much money, but that's largely gone away post-boom.
Beyond under-capitalization, you have to keep a clean captable-- you can't
issue tons of shares to everybody. More subtly, preferred shares have to
have reasonable preferences-- anti-dilution clauses, participation rights,
etc. all need to be managed carefully. If you expect to raise money at
ever-increasing valuations, don't expect the preferences to get gentler
with later rounds of financing. Down-rounds and dilutive stock splits are
morale-killers, if only because employees never have as large a stake as
they want ("deserve"). You risk poaching from competitors selling
less-diluted stock as well as employees comparing stakes (bad form, but it
happens). Later employees are invariably unhappy with the ongoing
contributions of some early employee who has a much larger stake,
nevermind that the earlier employee took outrageous risk in joining.
It's true that lots of other things can be fixed-- so my last piece of
advice is to get good corporate attorneys. One veteran's advice: your
lead attorney should be at least 40 years old, because it's mathematically
impossible to have enough experience otherwise. You also want attorneys
who believe in you and your business, and will give you attention and
priority. Treat your attorneys like board members, making sure to
reinforce the future potential of the business-- but also take steps to
make them look good (to friends, colleagues and family members) for
representing you. Top firms can have lame attorneys, but top attorneys
rarely work for schlocky firms. Don't be afraid to ask questions, and
check the answers with information on the internet.
- revenue management. [Apr'04] Once the company is viable, the
product is good enough to sell, production is under control, you have the
pricing structure and channels chosen-- then don't wait: bring in
professional revenue management, i.e. someone who will take responsibility
for "making the numbers" and who has a full bag of tricks for maximizing
revenue by the dollar, including incentive programs, cutting special deals
with key customers, bundling tricks -- but who also is also a responsible
manager over the long haul, who'll properly incentivize salespeople,
support staff and partners, who'll maintain price-consistency over
time, between products and between customers, and who'll maintain the
integrity to keep you out of jail.
The key is to stop and ask yourself, "why is this person asking me this? what
do they really want to know?" The remainder of this section are some common
questions you may be asked, likely scenarios for why it's being asked, and
some possibly-good answers. If you have ideas, please email me: asah @ midgard.net