Takealot replenishment: the simple reorder formula most sellers don't use

TL;DR. Most Takealot sellers reorder by gut feel. No formula, no data, no track record of what each SKU actually needs. The math behind reorder quantity is genuinely simple — daily sales rate × cycle length + safety buffer — and it works the same way for fast-movers and slow-movers (the velocity feeds the formula; the formula doesn't change per tier). What separates the disciplined sellers from the rest isn't sophistication; it's that they actually run the formula, every cycle, on every SKU. The compounding consequence of gut-feel reordering: intuition rounds to round numbers, anchors to last cycle's order quantity, and systematically under-weights changing demand. Replace gut feel with the formula — once per cycle, 30 minutes for a top-50-SKU catalogue — and the inventory mix self-corrects within 2–3 cycles.

Most Takealot sellers don't have a reorder formula at all

We surveyed how SA Takealot sellers actually decide reorder quantity. The dominant answer wasn't "trailing 30-day sales × cycle length", or "last cycle adjusted", or "MOQ-driven". It was "gut feel — no formula at all".

That's not a criticism of the sellers — most are building a business in a domain where the marketplace doesn't provide a tool for it, the right formula isn't widely taught, and "I'll just order what feels right" works well enough not to be catastrophic. Gut feel doesn't produce zero results; it produces suboptimal results, with a specific failure pattern:

• Intuition rounds to round numbers — "order 100" beats "order 87" in the seller's head, even when 87 is what the maths actually says
• Intuition anchors to last cycle — "I ordered 100 last time, this time 110 feels right" is the most common implicit logic, and it under-reacts to genuine demand shifts
• Intuition is too coarse to differentiate well between SKUs at scale — at 5 SKUs, gut feel is fine; at 50, three are over-ordered and three are under-ordered every cycle, and the seller doesn't know which

Stockouts flagged the trigger question — when to reorder. This post is the quantity question — how much to reorder when you do. Two different decisions on the same SKU, both of which gut feel handles poorly.

The reorder formula

Three inputs, one output:

Reorder quantity = (daily_sales_rate × cycle_length_days) + safety_buffer

• daily_sales_rate — units sold per day, computed from the trailing 30-day sales data (adjusted for trend if the SKU is moving)
• cycle_length_days — how many days the order needs to last before the next reorder cycle hits. Not just lead time — this is the time between reorder events, which is usually longer than lead time
• safety_buffer — extra days of cover above the cycle to absorb variance (7–14 days for stable SKUs, more for volatile or imported lines, per the stockouts post)

A worked example. A SKU sells 4 units/day on average. The seller's reorder cycle is 30 days (they place orders monthly). Safety buffer is 10 days.

Reorder quantity = (4 × 30) + (4 × 10) = 120 + 40 = 160 units

That's the number. Not 100 (gut feel). Not 150 (last cycle + a bit). One hundred and sixty.

If gut feel had said 200 ("better safe than sorry"), the seller's just put 40 extra units into DC storage burning fees for ~10 extra days. If gut feel had said 120 ("don't want too much cash tied up"), the seller's heading into a stockout 10 days before the next reorder lands.

The formula isn't magical. It's just defensible — the seller can point at the inputs and explain the output, which means they can also adjust each input deliberately when conditions change.

Why the same formula works for fast-movers and slow-movers

The intuition many sellers have is that different SKU types need different reorder logic — "A-class items need a different framework than C-class items". They don't.

The same formula applies. The velocity is an input; it's not something that changes which formula you use. A SKU selling 0.5 units/day computes its reorder quantity the same way as a SKU selling 50 units/day:

SKU	daily rate	cycle	buffer	Reorder quantity
Fast mover	50	30	10	(50 × 30) + (50 × 10) = 2,000 units
Mid-tier	4	30	10	(4 × 30) + (4 × 10) = 160 units
Slow mover	0.5	30	10	(0.5 × 30) + (0.5 × 10) = 20 units

The maths self-tiers. Fast movers naturally get larger orders; slow movers naturally get smaller. The formula handles it without you needing to maintain separate frameworks.

The only thing that might change per velocity tier is the cycle length. Some sellers run very-fast-movers on a 14-day cycle (smaller, more frequent orders to limit storage cost) and slow-movers on a 60-day cycle (larger, less frequent orders to amortise shipping). But the formula is the same; only the input changes.

The trending-up trap

Trailing 30-day sales is the right base — but only when demand is stable. For SKUs whose sales are climbing (or falling), trailing data systematically lags reality.

A SKU selling 2 units/day three months ago, 3 units/day two months ago, 4 units/day last month, and 5 units/day this month is on a clear upward trend. Trailing 30-day average gives you 5 units/day — already behind. By the time the next reorder lands, it'll likely be selling 6 or 7 units/day. The formula's daily rate input needs to reflect forward demand, not backward demand.

Practically, two adjustments:

• Trend multiplier: if month-over-month growth is +20%, multiply trailing daily rate by 1.2 to get forward rate
• Forward forecast: more rigorous — model the trend explicitly and project N days forward; use the projection as the daily rate input

The math doesn't have to be sophisticated to beat trailing-only. A simple multiplier captures most of the benefit. The killer mistake is sellers who don't adjust at all — they reorder against trailing data on every SKU and chronically under-order the climbers.

This is one area Gadjet's reporting automates: the daily rate per SKU is computed with trend adjustment built in, so the input to the reorder formula reflects forward demand by default. But the principle is what matters: forward rate, not trailing average, when the SKU's trajectory is non-flat.

Cash allocation — when you can't reorder everything

Most sellers don't have unlimited working capital. At some point, the formula says "reorder R150,000 of stock this month" and the bank balance says "you have R80,000". The decision becomes: which SKUs get the cash?

Three useful rules, in priority order:

1. High-revenue, high-margin SKUs first. The SKUs that pay the bills are the ones you can't afford to stock out. Cash here protects existing revenue.
2. Fast-velocity SKUs over slow-velocity ones. A R10,000 order on a fast-mover turns over in 30 days; the same R10,000 on a slow-mover sits for 6 months. Faster turnover = more revenue per rand of working capital.
3. Profitable SKUs over breakeven SKUs. Run the unit economics on each SKU. Some "selling well" SKUs are actually losing money after Success Fee + fulfilment + storage. Don't fund those at the expense of profitable ones.

What this looks like in practice: the catalogue gets ranked by revenue × margin per SKU. Top of the rank list gets fully funded reorders; lower-ranked SKUs get partial orders or skipped this cycle. The ranking is mechanical; the discipline is in following it instead of falling back to "reorder a bit of everything".

The 30-minute monthly habit

The whole replenishment loop, end-to-end:

1. Pull trailing 30-day sales per SKU from the Seller Portal
2. Compute daily sales rate — total sold ÷ 30
3. Apply trend adjustment if the SKU is climbing or falling (compare last 7-day rate to trailing 30-day rate; if they differ by more than 20%, adjust)
4. Apply the formula: reorder qty = (daily rate × cycle length) + safety buffer
5. Rank SKUs by revenue × margin if cash-constrained; allocate working capital top-down
6. Place the orders with suppliers; update your tracking sheet

For a top-50-SKU catalogue, this takes about 30 minutes the first time and ~15 minutes once you've built the spreadsheet template. The compounding return on those 15 minutes per month: stockouts drop, storage fees on slow-movers drop (because you're not over-ordering them), and the inventory mix self-corrects toward your actual velocity profile within 2–3 cycles.

Compare against gut feel: every reorder cycle introduces fresh variance. Some SKUs over-stocked, some under. The maths never settles. The seller works harder because the inventory keeps generating new problems.

What the formula doesn't tell you

Three things the reorder formula deliberately doesn't include:

1. Promotional events. Black Friday, Daily Deals, school terms — these change the cycle's expected daily rate. Either pre-position stock specifically for the event (separate calculation) or temporarily multiply the daily rate during the promo window.
2. Supplier MOQs and shipping economics. If the formula says "order 73 units" but your supplier's MOQ is 100, you order 100. If shipping is cheaper at 500 units than 200, you might bias up. The formula tells you the demand-driven number; supplier-side constraints can adjust upward but rarely should adjust downward.
3. Risk of supplier discontinuation. If you suspect your supplier might stop carrying a SKU, a one-time over-order makes sense — but treat it as an explicit hedge decision, not "the formula".

These are seller-judgement overlays on top of the formula. The formula gives you the starting number; you apply the overlay deliberately, not as an excuse to fall back to gut feel.

Frequently asked questions

Why is daily sales rate computed over 30 days, not 7 or 14?

Thirty days is enough to smooth out weekday/weekend variance and most non-event noise. Seven days is too short — a single high-sales day distorts the rate. Sixty days is too long — by the time you're computing it, the data is stale. The exception: very-fast-moving SKUs benefit from a 14-day rate because demand can shift faster than the 30-day average reflects.

What if the SKU is new and there's no 30 days of sales history?

Use whatever history exists, weighted toward more recent days. For a 10-day-old SKU, compute the rate from those 10 days but add a wider safety buffer to absorb the higher variance. After 30 days, switch to the standard formula. New-SKU reorders are always a guess; widen the buffer to compensate.

How do I pick the right cycle length?

Cycle length is mostly a function of cash and supplier logistics. Shorter cycle = smaller orders, lower storage cost, more orders per year (more admin overhead). Longer cycle = bigger orders, better shipping economics, more inventory tied up. For most SA Takealot sellers with local supply, monthly is the right starting cycle; tighten to 2-weekly on hero SKUs if storage is biting, lengthen to bimonthly on stable slow-movers.

Should I use weighted moving averages or other forecasting methods?

For most catalogues, simple is enough. A weighted average that prioritises recent days adds 5–10% accuracy over flat average. Exponential smoothing adds another 5%. The formula's accuracy ceiling on Takealot data is well below the impact of just doing it consistently — sophistication beyond a trend multiplier delivers diminishing returns. Spend the saved effort on the SKUs themselves.

How do I handle bundled SKUs in the reorder math?

The bundle consumes underlying inventory. If your "Pack of 5" sells 2 bundles/day, that's 10 underlying units/day going out. The reorder formula on the underlying unit must include both the singles and the bundles. Practically: track underlying units, not listing units, in the cycle.

What's the relationship between this formula and the safety buffer from the Stockouts post?

The same safety buffer concept. The Stockouts post framed the buffer as days-of-cover above lead time (when to reorder); this post frames it as days-of-cover above cycle length (how much to reorder). Same number, two different uses. A SKU with a 10-day safety buffer triggers a reorder when stock drops to 10 days + lead time, AND each reorder includes 10 days of buffer on top of the cycle's expected consumption.

Can I automate this?

Yes. Gadjet's replen-plan skill computes daily sales rate with trend adjustment per SKU, applies the formula with your configured cycle length and safety buffer, ranks by revenue × margin for cash-constrained decisions, and produces a single ranked reorder list. The point isn't replacing your judgement on each SKU — it's collapsing the 30-minute monthly spreadsheet into a one-screen review where you confirm or adjust the numbers. See Gadjet.

What to do this week

If you've never run the reorder formula systematically:

1. Pick your top 20 SKUs by revenue contribution
2. Pull 30-day sales for each — compute daily sales rate
3. For each SKU, check the 7-day rate vs 30-day rate. If 7-day is materially higher, the SKU is trending — apply a 1.2× to 1.5× multiplier to the daily rate
4. Decide your cycle length for the catalogue — start with 30 days unless you have a specific reason otherwise
5. Apply the formula: (adjusted daily rate × cycle) + safety buffer (start at 10 days)
6. Compare against your next planned order quantity for each SKU
7. Note where the formula and your plan diverge — for each gap, decide which is right (and adjust the gut feel for next cycle)

One cycle, one formula, one habit. The reorder mix usually stabilises around your real demand profile within 2–3 monthly cycles. The intuition that took years to build gets replaced with a 30-minute monthly process that produces better results — and survives the seller going on holiday.