What Do You Do When Six Months of Stock Sells in Three Weeks? Seedlip's Capacity Reset
Short answer: Do not multiply the next order by the sellout speed. First separate repeatable weekly demand from launch publicity and scarcity, then raise capacity one tested step while protecting specification, cash and delivery. Fund the next run plus one failure scenario, and cap customer promises at confirmed output. A sellout proves that one quantity sold; it does not prove how much the market will absorb after novelty fades.
Selling out feels like the easiest business problem. It becomes dangerous when excitement authorises a production commitment larger than the evidence.
The next batch ties up more cash, introduces new people and processes, and creates promises to retailers and customers. Scaling is a quality and working-capital decision before it is a celebration.
Use the Capacity Shock Ladder
The Capacity Shock Ladder has five rungs. Climb only when the evidence on the current rung supports the additional fixed or committed capacity above it.
| Rung | Decision | Evidence required | Main failure | |---|---|---|---| | 1. Verify demand | Estimate normal weekly orders | Source, price and repeat data | Launch spike treated as baseline | | 2. Stabilise process | Reproduce the accepted product | Batch records and quality pass | Larger run changes the product | | 3. Fund the gap | Pay production before collection | Dated downside cash forecast | Growth exhausts working capital | | 4. Add capacity | Extend shifts, partner or equipment | Cost per good unit and lead time | Capacity added at wrong bottleneck | | 5. Release promises | Accept orders within confirmed output | Allocated stock and service plan | Overselling damages trust |
My view is that a young product business should ration a genuine sellout rather than accept unlimited back-orders. Scarcity loses some near-term revenue. Broken delivery and inconsistent product can lose the category you just created.
What Seedlip's first batches revealed
Ben Branson developed Seedlip for about two years before launching its first 1,000 bottles through Selfridges. He told Vogue that he expected the batch to last around five months, but it sold in three weeks. He made another 1,000, which sold in days, and a later 1,000-bottle online release sold in about 30 minutes. The same founder interview says a batch took roughly six weeks.
Branson's later account of the capacity change says he moved from 1,000-bottle batches to roughly 7,000 or 8,000 bottles within three months and outsourced production. The successive sellouts made the capacity constraint visible, but they did not prove that the larger batch would sell at the same pace.
Those figures are founder-reported and do not disclose price, contribution, retailer terms or the cost of outsourcing. They show the operational question clearly: demand arrived faster than the production cycle could respond, so capacity and founder-made production had to change together.
Remove the launch spike from your forecast
Reconstruct the sellout by day, channel and source. Separate customers reached by launch press, personal contacts, retailer placement, organic search, referrals and repeat purchase. Record full-price units and promotional units separately.
Calculate three demand rates:
- Peak rate during the strongest publicity window.
- Settled rate after that window closes.
- Repeat-supported rate from customers or accounts ordering again.
Plan the next run from the settled rate, then use the repeat rate to justify further capacity. Peak demand is useful for queue and fulfilment planning, not as the sole basis for annual production.
If the item remains unavailable, use a dated reservation or waiting list rather than treating every email as an order. Ask customers to reconfirm when stock is ready. A paid pre-order is stronger evidence but creates refund, delivery and consumer-law obligations. Obtain qualified advice for your terms and market.
Freeze the specification before changing the process
Write the accepted product and pack specification, critical process settings, tests and tolerances. Retain approved samples and batch records. The new producer or larger line should demonstrate the same outcome before making the full run.
Scale-up can change mixing, heating, settling, fill, seal, finish and shelf behaviour. A formula that works in 100 litres may not behave identically in 1,000. Approve a technical-transfer batch, then a limited commercial batch, before committing maximum volume.
Assign who releases product, handles deviations and authorises any ingredient, material or process change. A contract manufacturer provides capacity, not automatic quality control. Verify its certifications, insurance, traceability, recall process, financial position and available line time.
Food, drink, alcohol, labelling and product-safety duties vary by jurisdiction. Use qualified technical and legal advisers. A non-alcoholic positioning does not remove every alcohol, food or marketing rule.
Find the real bottleneck
Map output from raw material receipt to delivered order. Measure the maximum good units per week at each stage: production, quality release, packaging, warehousing, picking and carrier collection. The lowest reliable number is capacity.
Do not buy filling equipment when label application or laboratory release is limiting output. Do not outsource manufacturing when a component with a 12-week lead time is the constraint.
Compare four responses:
| Response | Best when | Risk | |---|---|---| | Additional shift | Equipment has spare hours | Labour, supervision and fatigue | | Larger batch at current site | Process is stable and equipment fits | Concentrated quality loss | | Contract producer | Specialist capacity exists | Transfer, minimums and dependence | | New equipment | Demand is durable and process known | Capital, installation and utilisation |
Choose the smallest reversible step that clears the measured bottleneck.
Worked example: Still North Aperitif
Still North sells 900 bottles from a launch batch in three weeks at £24 each. Two hundred units came from a launch event, 280 from press coverage and 420 from ordinary website and trade orders. Only 110 customers have had time to reorder.
After the launch coverage ends, confirmed weekly demand settles at 115 bottles. Management models 16 weeks until the next review, adding a 20% service buffer:
115 bottles x 16 weeks x 1.20 = 2,208 bottles
It rounds the next run to 2,200, not the 4,800 suggested by annualising the three-week launch rate.
| Cash item | Calculation | Amount | |---|---:|---:| | Production and pack | 2,200 x £8.40 | £18,480 | | Freight and quality testing | Given | £2,650 | | Launch receivables still unpaid | Given | £6,200 | | Operating cash for eight weeks | Given | £14,000 | | Cash needed before collection | Total | £41,330 |
Still North has £47,000 available, leaving only £5,670. A failed 10% batch would cost 220 x £8.40 = £1,848, but a four-week retailer delay could add £9,000 of exposure. The proposed run does not leave enough downside cash.
The company therefore orders 1,600 bottles, costing 1,600 x £8.40 + £2,650 = £16,090, and caps retailer allocations. It uses a technical-transfer run to confirm output before authorising the next 600. The smaller order may sell out again, but it keeps the business able to replace a failed batch and pay normal commitments.
These figures are illustrative. Use your own tax, contract, shelf-life and payment data.
Allocate scarce stock deliberately
Do not reward the loudest buyer automatically. Protect the channels that produce repeat evidence, pay predictably and position the product correctly. Give each confirmed account an allocation, delivery date and reorder condition.
Direct customers may provide faster cash and richer feedback. A respected retailer or hospitality account may provide discovery and credibility. Compare contribution and strategic learning rather than revenue alone.
Avoid spreading tiny quantities across too many accounts. Empty shelves and unreliable replenishment frustrate retailers. A narrower group supplied consistently can teach more than national availability that disappears in a day.
Communicate stock limits accurately. Do not manufacture scarcity or imply that production is imminent when supplier capacity is unconfirmed.
Reset capacity over the next 14 days
In the first three days, split launch orders by source, date, price and repeat status. Map the current production cycle and identify the lowest reliable weekly output.
During week one, freeze the accepted specification and obtain written capacity, cost and lead-time proposals. In week two, build base and downside cash cases, including a failed batch and late retailer payment. Authorise only the next ladder rung, allocate its output and set the evidence required before another increase.
Related guides
Frequently asked questions
How much more stock should I order after selling out?
Order from settled weekly demand over the next review period, plus a service buffer you can finance, rather than multiplying the launch rate. Remove event, press, discount and scarcity spikes. Add confirmed trade orders and repeat demand, then test a downside where sales are 30% lower or payment is delayed. The right number also depends on shelf life, supplier minimums and replenishment lead time. A second sellout can be less damaging than unsold or defective stock that consumes all cash. Increase again after the new run reproduces quality and demand.
Should I accept pre-orders while stock is unavailable?
Only when the production slot, realistic delivery window, refund process and customer communication are reliable. Pre-orders provide stronger demand evidence and cash, but that money comes with an obligation. Keep enough available to refund a delay or failure and do not use it for unrelated spending. Consumer, payment, tax and insolvency rules vary by market, so have the terms reviewed. If timing remains uncertain, a no-payment reservation with reconfirmation may be safer. Never advertise an estimated date as guaranteed when the producer has not confirmed it.
When should I use a contract manufacturer?
Use one when the process is sufficiently defined, the partner has verified capability and the total cost is better than your realistic in-house alternative at required volume. Include technical transfer, minimum runs, tests, freight, wastage, supervision and dependency. Outsourcing is not a cure for an unstable formula or unclear specification. Run a transfer batch and verify traceability, insurance, capacity and corrective-action procedures. Protect formulas, tooling and continuity contractually. Sector and product rules may leave you responsible even when another company makes the goods.
How can I tell whether scarcity caused the demand?
Compare behaviour after stock returns. Measure full-price conversion, cancellations, repeat purchase and the share of reservations that reconfirm. Track whether customers buy because the product solves the same job or because a launch story created urgency. Remove artificial countdowns and broad publicity from one controlled release. Trade reorders are useful when they reflect consumer sell-through rather than stockpiling. Scarcity can accelerate a real decision, but it cannot sustain repeat demand. Forecast from settled sales after availability normalises, not from the fastest sellout interval.
Should I raise the price when demand exceeds supply?
Test price only after confirming that capacity, not a temporary launch spike, is the constraint. A higher price can ration demand and fund capacity, but it may change the customer and channel position. Quote the higher price to a limited, comparable cohort and measure contribution, conversion and repeat purchase. Do not exploit customers who have already paid or change agreed retailer terms unilaterally. If the product is underpriced, correct it deliberately. If the bottleneck is a short delay, a transparent allocation may preserve more trust than opportunistic pricing.
What if the larger batch fails quality checks?
Quarantine it, stop dispatch and follow the documented deviation, recall and notification process. Do not blend, relabel or sell non-conforming stock without qualified technical and legal approval. Investigate whether the cause is specification, scale, input, equipment or operator, then prove the corrective action on a bounded run. This is why correction cash and traceable batches are required before scaling. Insurance may not cover every loss. Food, drink and consumer-product obligations vary, so involve the appropriate competent professionals and authorities promptly where safety or compliance is affected.
When is it safe to buy production equipment?
Buy when durable demand will use the capacity, the process is stable, the equipment clears the actual bottleneck and ownership beats outsourcing after installation, maintenance, labour, downtime and finance. Model utilisation under a downside sales case. A machine operating at 25% can raise unit cost despite a lower theoretical production rate. Confirm premises, power, safety, training and approval requirements before paying a deposit. Lease or staged contract capacity may preserve flexibility while demand settles. Equipment should follow process knowledge, not the emotional high of a sellout.
Comments (0)