Thursday, December 22, 2011

Bottom Up

In a board meeting the other day, one of the board members gave an hour presentation about the company's prospects.

I have four problems with the presentation:
  • the member has no operating experience of any kind (MBA--all hat and no cowboy)
  • it's an hour of my life I won't get back
  • it's the CEO's role; it's good when board members contribute but this was over the top and very little of it was relevant
  • the projections were based on top-down analysis.
A top-down analysis is tempting. You identify a market size. You then stake a claim on part of that market, oh, let's say 1%. 

Really. I was waiting for the punch line, but none came. From there you rationalize that with some simple division based on a hypothetical average selling price to get your target. "If we only add 1000 customers at $1,000 each, we'll have a million more of revenue". 

Priceless. 

So what's wrong with 1%? 

1% tells you nothing about customers. Nothing about product. And nothing about how to get there. Worse, it almost says you're willing to only take 1% of your target. 

That's where Bottom Up comes in. 

Bottom up means to break your target down to its component parts: number of leads, number of closed customers, average sale, etc, over a short period of time, and the marketing and sales activities you need to achieve those assumptions. 

The sales and marketing work should be expressed as resources--the spend on each marketing activity (online ads, adwords, SEO, whatever) and sales effort (number of calls per rep, number of reps, number of emails, responses to inquiries, etc. You can automate but early on you'll have difficulty converting without customer contact).  

You should validate those assumptions through direct contact with your market: call them. Methodically call and have conversations. Take lots of notes. 

Then develop a set of questions to ask to the broader market. Test your assumptions. Don't vary the questions, unless it seems you're way off on the initial set. 

Interview enough prospects to be able to reasonably quantify the results. A statistically significant sample is 31, but that's with a homogenous group, and I'm going to guess that there's wide variation within the first 31 people you talk with because you haven't learned how to narrow your market. 

So shoot for 300 completed interviews. It's a lot of work but it's the most valuable work early on. 

Out of those, you'll find some very willing to keep in touch with you and try your stuff, partly because what you're doing is much more exciting than their daily work, and partly because you're just so awesome and amazing that they're happy to help you along to greatness. 

Let's say you spend 10 minutes per prospect. That's 3,000 minutes, or 50 hours. Not bad, really. But it won't really work that way--it will likely take two to three times that long because you won't connect on every call, and some will last much longer (sometimes because you're simply enjoying it, and others because they are enjoying it. I once had an hour call with a woman who was simply depressed about her work and life...).

Add in a week of interviews, and another week of following up with those clearly interested. Don't waste their time; ask a specific question to each of them. Be appreciative and gracious.

The good news? You'll have a great set of data on which you can base your bottom-up assumptions. You'll have a much better sense of your market and mission. Your gut will truly be informed intuition. 

Your questions should be about the problem you think you're solving. Price should be in there--how willing would you be to pay $100/month for something that solves the problem? $75/month? $1000/year? $10 one-time?

Pricing questions can be tough, so I tend to start high and work down from there. I find pricing to be one of the toughest things to set. You have to price against competitors, budgets, value perception, biases, etc. I highly recommend spending some time researching pricing methodologies until you get comfortable--this isn't easy stuff, and blowing your pricing could leave money on the table or scare prospects away. 

So now that you have your research on which you can base your assumptions, you can start building your model. You might want to close 100 customers per month in your top-down 1% faulty logic, but the data suggests your close rate will be 10%, which means you need 1000 qualified leads and the team and tactics to close them. 

Do you need a sales force? 

Early on, it's just you--everyone should know how to close a sale. Is it an automated process? Good luck with that. Early on, you're going to have to have some contact with early prospects so they are comfortable parting with their money. 

I've left out a lot of stuff. Steve Blank has a through set of posts (and a book) on customer development. You should buy the book or read through his posts--it's the most relevant, actionable stuff out there. 

It's easy to create assumptions out of thin air. Don't. 

Do the work. Know your market and prospects from the bottom up. 1% of bullshit is still bullshit. Bottom up analysis is the real deal, and you can only nail it by talking with a lot of prospects. 

UPDATE: Brett Topche notes in the comments that top-down is still useful as a reality check on market size. He's right: you need both top-down and bottom-up approaches to nail your projections. 

No comments:

Post a Comment