How to Change the World: The Art of Projections in a Dotcom 2.0 World

The world is running amok with entrepreneurs pitching every sort of Web 2.0, social networking, user-generated-content startup. It’s the attack of the bull-shiitake startup projections, so I’m losing my hearing; there’s a ringing in my head, and I get dizzy every once in a while. Before the world implodes (again), here is a top-tenish list of ways to create realistic projections in this Dotcom 2.0 world.

Under-promise and over-deliver. I have never seen a company meet or exceed its initial forecast. Entrepreneurs come up with numbers that they guess investors want to hear, and everything goes down hill from there. As a rule of thumb, dividing sales forecasts by one hundred and adding one year to the projected shipping date is about right for startups without a prototype. For startup with a prototype, dividing by ten and adding six months is about right.
Forecast from the bottom up. Figure out how many business development and sales meetings you can get per week--that is, four or five. Then multiply this number by the percentage that will be successful. Then add six months to close the deal. This forecasting method yields a much smaller number than the “conservative” method of assuming that you can get at least one percent market share.

Once you’re done with the plan, show it to the rest of the management team and demand honest feedback. This is the only way to make a bottom-up plan truly bottom up. Don’t let anyone--for example an ego-maniac chairman--make the company sign up for a plan that isn’t achievable with at least eighty percent certainty.
Don’t go too far out: twelve to eighteen months is the maximum. Anything beyond that is a waste of everyone’s time because you really don’t know when you’ll ship, and you can only fantasize about customer adoption. If you’re into five-year forecasts, go to work for a nice consumer packaged-goods company that’s been around for fifty years.
Plan to re-forecast every three months. Otherwise, forecasting is a joke: You get approval for an annual budget and then re-forecast it in the following board meeting. It’s better to know that re-forecasting is necessary once a quarter than to pretend that “this time we got it right.”

However, there is a danger in the rolling three-month forecast: Employees will start to believe that “investors don’t mind” constant shortfall (I hope you’re not this clueless). In a startup, everything is “near term.” The long term for a startup is a year--get used to this mentality.
Don’t let costs get in front of revenue. I know, I know: Your startup is going to be the fastest growing company in history, so you need to build an infrastructure to support the onslaught of customers. Dream on. Always run leaner than you think is necessary because your challenge will be creating demand, not fulfilling it.

Specifically, keep your net burn under $250,000 per month. How can I pick a number like this out of the air? On the other hand, what good does a vague answer do? You don’t have to believe me, but $250,000/month is the magic number for a “typical” venture-capitalist deal. Anything above this, and you’re probably throwing money away building infrastructure for non-existent customers. Just once in my career, I would like to hear, “We ran too lean and sacrificed growth” instead of what I hear all the time: “Costs got ahead of revenue, so we need to cut back but we’d prefer you tell us we don’t have to because it will harm morale.”

If the $250,000 guideline isn’t appropriate, then at least do a sanity check. Look for insane assumptions like achieving the fastest growth of any company in history and doubling your salesforce in a month. It’s tactical and practical to think like the salesperson on the street trying to make quota and an HR manager trying to fill positions because these functions are easy in Excel but hell in reality.
Collaborate with your investors. It’s just plain dumb to show your Holy Grail of a forecast to investors for the first time at a board meeting. You should feel them out in advance, and never be in the position of guessing what you think they want to see. Collaboration is especially important if you have bad news. Surprising venture capitalists with good news is never a problem.
Think in terms of per-unit profitability. It may be acceptable to lose money on every unit for a time, but at some point you have to make money on every unit. And don’t count on Google to buy you out because “getting lucky” is not a viable strategy. Also, you need to know exactly how much you’re losing on every unit so that you can measure progress towards profitability.
Plan for marketing costs. Don’t depend on wishful-thinking marketing based on virality, buzz, TechCrunch, and a demo at Demo. It’s true that some companies do achieve success this way, but we’ve heard of them because they are few and far between. To use a sports analogy, we all know who Michael Jordan and Wayne Gretzky are because they are rare examples, not because their story is common.

You need to explain your demand-creation process in a mechanical, not magical, way: ad rates, click through rates, unique visitors per month, conversion rates, revenue per customer, etc. Ultimately, the underlying assumptions in your marketing model is the key to the fundability and viability of your startup. “We’ll get it on TechCrunch and then viral marketing will be easy because we have such a compelling product” doesn’t cut it.
Create a one-page report and stick to it. It seems like thirty minutes of every board meeting is spent explaining a new way to report revenues, costs, and metrics. You would think that you could pick a few numbers that indicate what’s happening at the startup and see the historical trends--but not if you change reporting methods every month. One innovative way to fix this might be to reduce the CEO’s and CFO’s stock options by ten percent every time they change the report.

You’ll impress investors if you present your projections and your results in the same format. For example, if you use QuickBooks categories for your general ledger, then use these same categories for your projections. The good news is that your numbers will be much easier to understand; the bad news is that you’ll be a lot more accountable. :-)

Finally, you’ll astound investors if you show up with what’s called a “Waterfall Forecast,” a report that shows how your forecast has changed over time. (Full credit to Josh Kopelman for blogging about this.) I would add one more calculation to Josh’s model: showing the variance between actuals and projections so that the depth of your fantasies (or achievements) is more noticeable.
Never miss a cost projection. It’s relatively okay to miss a revenue projection because forecasting for a startup is truly a crapshoot. If you miss a cost projection, however, then you’re entering the realm of cluelessness. There is no excuse for it, barring an act of God like a factory burning down that produces the raw material that you need. Even then, you should have had a backup source. You should be able to come up with numbers like twenty percent for payroll taxes; $500/month for employee benefits; $3000/employee for equipment costs; and $25/square foot per year for rent.
Think big. Build on small successes at first, but if you go five years out, find a way to get to $100 million in sales. If you can’t imagine this level of sales without being high on crack, then you should face the facts: your company probably isn’t a “VC deal.” It may be a perfectly viable business, but it probably isn’t one for venture capital. The only way you’ll get to this five-year point is hand-to-hand combat with a ninety-day outlook at the start, but it’s important for everyone that a pot of gold can be at the end of the rainbow.

While writing this blog entry, I referred to a book by Bob Prosen called Kiss Theory Goodbye: Five Proven Ways to Get Extraordinary Results In Any Company, which helped me considerably.