Logistics have a magic when you’ve acquired an intuitive sense for numbers and constraints. Most of yesterday was spent on the phone with Beth, talking out the various spreadsheets and plans we’ve got going. My own number goals are somewhat exacting: covers all constraints, easy to remember, easy to combine and adjust in my head, feels fair and affordable.

Finding workable numbers is a real skill which is surprisingly similar to my chief talent: designing software. Lots of conflicting constraints, lots of stuff to keep in mind at once, lots of interdependencies … all boiled down into something simple and workable.

As an example, Beth and I were trying to determine how much we should charge a location for Waveplace support over time. The goal is to find the smallest number that covers our costs. So starts a string of interdependencies dropping from my brain while Beth mostly says, “Sure”, “Ok”, “Sounds good.”

Tim: “Let’s say we get all forty locations going full time next year, with 25 laptops apiece, that’s 1000 active laptops. Forty locations means 200 mentors, with 10 level 3 mentors overseeing all of them. How much time does it take to read reports and respond to 10 L3 mentors. (Beth talks for a while about her experience and comes up with a 30 minute average per week per report.) Okay, 10 mentors means 1 hour per day just for L3 reports.

Tim: “Now how do we deal with those 1000 laptops? Assume complete communications, with our Q&A activity running, with Twitter-like updates, emails, possibly forum messages … call this the “response stream”. How much of this is generated per week? How much do we need to read and respond to ourselves? (We stumble around trying to imagine this a while.) Okay, let’s just pick a number. Out of 1000 active laptops, let’s say an average of 9% of them generate something we need to deal with. What’s the average time needed to deal with each? (More stumbling, which results in a 10 minute average) Okay, that’s 900 minutes, which is … 15 hours exactly. Add that to the L3 mentor time and you’ve got 20 hours a week supporting our locations … 4 hours a day.”

Tim: “Now let’s calculate our cost per hour. (Tim pulls up his spreadsheet and finds the monthly cost for Beth, including salary, health insurance, payroll taxes, etc.) Okay, divide this by days per month and hours per day and we get … just a bit over $25 an hour. This means support will cost us $500 a week, or $2000 a month. Since we’re supporting 40 locations, this means $50 a month per location.”

Tim: “Ok, looking at our per class monthly costs, we’ve got $600 for mentors and $600 for power/internet/food, which is $1200 a month or $14,400 a year. With support at $50 a month, that’s $600 a year, which added to $14,400 makes an even $15,000. Oh, great, now people are gonna think we just picked that number!”

Anyway, this is one example of dozens from the last few days. Picking numbers that are low enough but still cover costs is an art in itself, and such serendipitous moments like the one above (having it equal an exact $15,000) happen pretty often. Number sense is a talent. Not losing money from bad numbers is another.

Leave a Reply

Your email address will not be published. Required fields are marked *