In 1974, psychologists Amos Tversky and Daniel Kahneman rigged a wheel of fortune. It looked random, but it only ever stopped on 10 or 65. After each spin, they asked participants a question: what percentage of African nations belong to the United Nations? The wheel obviously had nothing to do with the answer, and everyone in the room knew it. Yet people who saw it land on 10 guessed about 25 percent. People who saw 65 guessed about 45. A meaningless number pulled their judgment straight toward it.
Tversky and Kahneman called this anchoring. Once we see a number, we stop reasoning from scratch. We adjust from the anchor, and we rarely adjust far enough. Now think about how most companies build a sales forecast.
Most forecasts are built backward. Leadership picks a number first: last year plus 15 percent, what the board expects, what the bank needs to see. Then the team spends weeks building a spreadsheet that arrives at it. The target comes first. The reasoning follows.
That’s the wheel of fortune with a fiscal calendar attached. The board, the bank, or last year spins it, and everything after is adjustment. It looks like analysis. It’s a story written around a number someone already picked.
Kahneman caught himself doing the same thing. Years later, he led a team of experts writing a high school curriculum and asked them how long it would take. About two years, they said. Then he asked a colleague how long similar teams had actually taken. Seven to ten years, the colleague admitted, and about 40 percent never finished at all. Kahneman’s team took eight.
The experts had a target. What they didn’t have was a count.
Esteemed member Greg Moser forecasts from the other direction. Greg is CEO of ShipCalm, a Carlsbad fulfillment company that warehouses, packs, and ships for e-commerce brands in beauty, apparel, supplements, food, and pet products.
When Greg plans the year, he doesn’t start with a revenue number. He starts with people: the customers who fit ShipCalm’s profile, some who already know the company and many who don’t.
Picture that audience standing outside a building with two doors.
Door One: someone who has never bought from you becomes a customer.
Door Two: an existing customer buys again, or buys more.
Outreach opens Door One. Retention and expansion open Door Two. Pricing and cross-selling can work on either.
The power of the idea is how much it rules out. A rebrand, a reorg, a sponsorship, a content push: each one has to answer a single question. Which door does this open, and for how many people? If the answer is neither, ask why you’re paying for it.
Greg borrowed the method from a CFO who once ran the books for a virtual world. In 2006, John Zdanowski joined Linden Lab, the company behind Second Life, as CFO. Second Life was a world its users built themselves. They bought land, developed virtual real estate, and traded in their own currency. Unusual business, same need for financial discipline.
Second Life sold islands. Every island brought in subscription revenue, and every island created work: servers, support, development. So Zdanowski tied headcount directly to the product. The company hired one person for every 130 islands sold. There were no arbitrary staffing budgets and no top-down targets. By his account, that discipline got the company to profitability within 18 months.
The principle is simple: find the unit of output, know what it earns and what it costs, and build the plan around how many units you can actually produce.
Zdanowski went on to co-found Weekly Accounting, a firm that provides outsourced accounting and fractional CFO work. The firm’s name is its thesis. Most companies close their books monthly, which gives them 12 chances a year to learn something. Weekly Accounting closes 52 times. As Zdanowski puts it: “If you’re not managing your business weekly, you’re managing your business weakly.”
His forecasting rules follow from that. Know your unit economics. Understand what actually drives revenue. Build a conservative budget you can hit every time. And if you keep missing it, don’t blame the market. Something is wrong with how you execute or how you’ve modeled the business.
A ground-up forecast is built from units of effort you can count and repeat. Take a trade show. Say one show gets you 50 real conversations with target accounts, those turn into 10 meetings, and the meetings turn into 5 new customers. Your numbers will differ, but the logic holds. One show, 5 customers. Two shows, 10. Three shows, 15.
Salespeople work the same way. If a fully ramped rep brings in 10 new accounts a month, two reps bring in 20. Every rep, show, campaign, and channel becomes a unit with a cost and an expected return.
That changes the question. A top-down forecast asks, “What number do we need?” A ground-up forecast asks, “What are we actually going to do, and what does each thing produce?” The first runs on hope. The second runs on your own conversion history.
It also changes the conversation with your board or your bank. Instead of “the market’s strong, so we’re projecting 20 percent,” you can say: two more trade shows, one new rep starting in Q2, and a retention program that gets customers reordering more often. Every dollar in the forecast has a reason next to it.
Once your efforts are units, strategy turns into comparison. Put your money where the return per dollar is highest.
Here’s a simple version. A $40,000 trade show that lands 5 accounts costs $8,000 per new customer. A $120,000 rep who lands 10 accounts a month once ramped costs about $1,000 per customer. A $20,000 retention campaign that gets 40 existing customers to reorder costs $500 per Door Two win.
Your numbers will be different. The point is that you can line them up side by side. Most budgets go to last year’s channels because they were in last year’s budget. Counting by the unit forces every channel to earn its place again.
A forecast built this way isn’t a guess. It’s a decision about where to spend, with the cost and the expected return written next to each line.
Simple multiplication eventually breaks, and a good forecast says where.
The audience runs out. The first pass through a channel finds the people who were ready to buy. Each pass after that reaches people who were less ready. No audience is infinite.
New capacity takes time. A rep who will eventually close 10 accounts a month won’t close 10 in month one. Plan as if they will, and you’ll be behind by March.
Delivery has a ceiling. Winning customers faster than you can onboard and serve them hurts quality, and poor quality slams Door Two shut.
A good model names these limits before you hit them.
This is the discipline behind Weekly Accounting’s approach and ShipCalm’s forecast. In the Esteemed MBAi we call it What Gets Measured Gets Done. Each driver becomes a number, and each number gets an owner: someone accountable for trade show yield, someone for ramping new reps, someone for repeat orders.
Back to that wheel. Left alone, we’ll build a story around whatever number we see first. The fix isn’t more willpower. It’s starting somewhere else: with the people outside your two doors and what it takes to bring them through.
A good forecast isn’t the biggest number. It’s the one you can explain, door by door.
Be Esteemed.