The Science of Demand: How to Predict Your Next Bestseller

Stop Guessing, Start Gaining: The Science of Demand

Stocking a retail store shouldn't be a game of "Vibes." If you have too much, your cash is trapped. If you have too little, your customers leave unhappy.

Mastering Demand Forecasting is about finding the "Goldilocks Zone"—just enough stock to maximize sales without tying up capital.

---

1. Calculating Sales Velocity

This is the heartbeat of your store.

• Formula: (Total Units Sold in 30 Days) / 30 = Daily Velocity.
• Zeneva Insight: Our AI calculates this automatically for every item. If your "Milo (Small)" has a velocity of 4.5 units, you know you need at least 32 units to survive a week.

---

2. Seasonal Peaks & Local Realities

Demand isn't a flat line. It waves.

• Holiday Planning: Zeneva's "Year-over-Year" reports help you see that your wine sales double in December.
• Payday Patterns: Most Nigerian retailers see a spike between the 25th and 5th of every month. Your forecasting should involve "Front-loading" stock just before these dates.

---

3. The Lead Time Calculation

Forecasting is useless if you don't factor in your supplier.

• If your supplier takes 5 days to deliver, and your velocity is 2 units a day, you must place your order when you still have 10 units left. This is your "Reorder Point."

With Zen AI, this entire process is automated. The system learns your patterns and tells you exactly what to buy, when to buy it, and who to buy it from.

---

4. Safety Stock: The Number Most Owners Skip

The reorder point above assumes your supplier always delivers in exactly five days. Nigerian supply chains do not work that way — fuel scarcity, port delays, a supplier's own stockout, or a truck that simply does not arrive.

Safety stock is the buffer that absorbs that variance:

Reorder point = (daily velocity × lead time) + safety stock

A workable starting rule: hold half your lead-time demand as safety stock for reliable suppliers, and a full lead time's worth for suppliers who have disappointed you before.

The important insight is that safety stock is the price you pay for supplier unreliability. If one supplier forces you to hold three weeks of buffer on a slow-moving item, that buffer is trapped cash — and the cheaper fix is usually a second supplier rather than more inventory. Quantifying the buffer turns a vague frustration into a number you can negotiate with.

---

5. Not All Items Deserve Equal Attention

The most common forecasting mistake is treating every SKU the same. A shop with 400 items cannot forecast 400 items carefully, so it forecasts all of them badly.

ABC analysis fixes this. Rank items by revenue contribution:

Class	Share of items	Share of revenue	How to manage
A	~20%	~80%	Track weekly, never stock out, negotiate hard with suppliers
B	~30%	~15%	Review monthly, standard reorder points
C	~50%	~5%	Review quarterly, accept occasional gaps, order in bulk to save effort

Your A items are where forecasting effort pays. A single stockout on an A item costs more than a month of C-item gaps. And a C item that has not sold in six months is not inventory — it is cash you spent, sitting on a shelf, that you could recover today by discounting it out.

---

6. The Trap: Forecasting Sales Instead of Demand

This is the error that quietly ruins otherwise good systems, and it is worth understanding precisely.

Your sales data records what you sold. It does not record what customers wanted. Those diverge every time you stock out.

Say you normally sell 30 units of an item a month. This month you ran out on day 12 and restocked on day 24, so you sold 15. Your system now sees 15, halves the velocity, and lowers the reorder point. Next month you order less, stock out sooner, and sell 10. The forecast falls again.

Within four months a strong seller has been forecast down to nothing — not because demand fell, but because the system was learning from its own failures. Owners then conclude the product "stopped selling."

Two defences:

1. Exclude stockout periods from velocity. If the item was unavailable for 12 of 30 days, divide by 18, not 30. Velocity is 15/18 = 0.83/day, not 0.5.
2. Record lost sales. When a customer asks for something you do not have, note it. This feels tedious and it is the only direct measure of demand you will ever have. Even a paper tally by the counter for one month is revealing.

---

7. Reading the Nigerian Calendar

Generic forecasting advice assumes a Western retail calendar. Yours is different, and these patterns are stable enough to plan around:

• The payday window (25th–5th). The most reliable pattern in Nigerian retail. Stock should peak just before the 25th, not after.
• School resumption (January, April/May, September). Stationery, uniforms, provisions, lunch items. Parents buy in a compressed window and buy everything at once.
• December. Not a uniform lift — beverages, confectionery, cosmetics and gift items spike hard while ordinary staples flatten. Forecast by category, not store-wide.
• Ramadan and Eid. Timing shifts about eleven days earlier each year, so last year's dates are wrong this year. Dates, beverages and provisions move sharply.
• Rainy season (roughly April–October). Foot traffic drops on heavy-rain days, which distorts weekly averages if you do not account for it.
• Salary delays. In months when public sector salaries are late, the payday spike moves rather than disappears. If your sales look wrong for a week, check whether the money simply arrived later.

---

Where to Start This Week

You do not need a full system to begin. In order of return:

1. Pick your top 20 items by revenue. These are your A class. Everything else can wait.
2. Calculate velocity for those 20 only, excluding any days they were out of stock.
3. Write down each supplier's actual lead time — the real one, not the promised one. Look at your last three orders.
4. Compute reorder points and put them somewhere visible.
5. Start a lost-sales tally at the counter for one month.

That is an afternoon's work and it will do more for your cash position than any software purchase, because it tells you which items are worth automating.

For the mechanics of acting on these numbers, see advanced inventory tips and ten ways to improve cash flow — the second is directly relevant, since over-forecasting is one of the most common ways Nigerian retailers trap working capital in stock.