Did Snag's Facebook Ads Reveal Real Demand or Just a Launch Spike?
Short answer: Judge Facebook ads by customer cohort contribution, not the platform's reported return. Track cash collected after returns, product and fulfilment cost, first-order acquisition cost, 60-day repeat contribution and results as audiences broaden. Call demand durable only when at least three consecutive cohorts repay acquisition within your cash window without relying on one creative, founder audience or discount.
Launch campaigns often reach the easiest buyers first. Novelty, founder contacts and one strong advert can produce a profitable week that cannot absorb the next £10,000.
The real question is not whether Facebook generated orders. It is whether the channel repeatedly finds customers who keep enough product and return enough contribution before the business must pay for stock again.
What Snag's reported growth can and cannot show
Snag founder Brie Read built the hosiery brand around a differentiated sizing approach and representative customer imagery. A detailed account of Snag's early business describes Facebook advertising as an important launch channel. In a University of Edinburgh Business School interview, Read says the business started with £100,000 raised from family and friends and handled hundreds of customer comments daily. Although that profile calls the growth "without outside investment", it does not mean no external funding: the founder explicitly describes the initial raise.
Public profiles report rapid revenue growth, but they do not disclose launch spend, acquisition cost, return rate, repeat contribution or creative-level cohorts. Revenue alone cannot establish that the ads were durable.
My position is that paid social should be treated as paid market research until a cohort repays its acquisition cost. A dashboard showing four times revenue is not evidence when returns, tax, stock and later discounts sit elsewhere.
Use the Paid-Demand Durability Matrix
The Paid-Demand Durability Matrix compares four economic signals across four sources.
| Source | First-order contribution | Returns | Repeat contribution | Scale behaviour | |---|---:|---:|---:|---| | Founder or existing audience | measure separately | by product and size | within defined window | Usually limited | | Narrow high-intent audience | after ad cost | by cohort | within 60 or 90 days | May saturate quickly | | Broader prospecting audience | after ad cost | by cohort | within same window | Main durability test | | Retargeting | avoid double counting | by original source | attribute carefully | Depends on prospecting supply |
Do not blend the rows. Retargeting often looks efficient because another channel created the visit. Founder audiences can produce unusually high trust. Broad prospecting tells you whether the offer travels.
Rebuild the first-order economics outside the platform
Start with settled revenue after cancellations and returns. Subtract product, packaging, picking, delivery subsidy, payment fees and variable support. What remains is contribution before advertising.
Maximum first-order acquisition cost equals that contribution if you are willing to earn nothing on the first order. A stricter business may require immediate surplus. A repeat-led model can tolerate a higher cost only when observed cohorts repay it within an affordable period.
Platform-reported revenue may include tax, cancelled purchases or attribution from people who would have bought anyway. Reconcile it with orders and payments before making budget decisions.
Follow cohorts, not calendar totals
Group customers by first purchase week and source. For each cohort, record:
| Cohort measure | Why it matters | |---|---| | Cash revenue retained | Removes cancellations and returns | | Product and fulfilment cost | Reveals actual first-order contribution | | Acquisition spend | Shows price paid for the cohort | | Second purchase timing | Determines cash payback | | Repeat contribution | Tests product value beyond the advert | | Complaints and exchanges | Detects creative or fit mismatch |
A good month can contain a worsening new-customer cohort and valuable older repeats. Calendar revenue hides that distinction.
Wait for the normal return period before declaring first-order success. For hosiery or clothing, size and quality issues can emerge after delivery. Attribute replacements and refunds to the acquiring cohort.
Test the promise as well as the picture
Advertising creative makes a claim about who the product is for and what it changes. High click-through with poor retained purchases means the promise attracts attention but the product page, price or product does not complete it.
Change one meaningful variable at a time: customer problem, proof, visual representation or offer. Keep a simple record of spend, audience, dates and product availability. Avoid judging a creative after a handful of purchases, but cap the loss before the sample grows.
Representative customer imagery can improve recognition and trust, but do not assume the mechanism. Compare conversion, return reasons and repeat behaviour for buyers acquired through each message.
Worked example: CurveFit Hosiery
CurveFit Hosiery is a fictional business; these figures are illustrative. It spends £6,000 over four weeks. Its average order is £38. Product, packing, delivery and payment cost total £17.40 for a retained order, leaving £20.60 contribution before advertising.
| Audience | Spend | New retained customers | CAC | First-order contribution after ads | |---|---:|---:|---:|---:| | Existing followers | £800 | 80 | £10.00 | 80 × £20.60 - £800 = £848 | | Narrow lookalike | £2,000 | 125 | £16.00 | 125 × £20.60 - £2,000 = £575 | | Broad prospecting | £3,200 | 128 | £25.00 | 128 × £20.60 - £3,200 = -£563.20 |
The blended CAC is £6,000 ÷ 333 = £18.02, which appears below £20.60. Blending suggests success. The broad cohort loses money on the first order.
Within 60 days, retained customers repeat at these rates: 30% existing, 24% narrow and 15% broad. A repeat order contributes £18 after fulfilment.
Broad repeat contribution is 128 × 15% × £18 = £345.60. Its 60-day result is -£563.20 + £345.60 = -£217.60. It has not paid back.
Narrow repeat contribution is 125 × 24% × £18 = £540. Its 60-day result becomes £575 + £540 = £1,115.
CurveFit should not double the whole account from the blended result. It can continue the narrow test while changing the broad message or product-page match. The existing audience is profitable but finite, so it cannot carry the scaling claim.
Test saturation deliberately
Increase spend in controlled steps, such as 20% to 30%, and hold long enough to cover the normal buying cycle. Watch CAC, frequency, conversion and cohort quality. A rising frequency with declining retained conversion suggests the audience has seen the offer too often.
Do not reset learning by changing budget, creative, landing page and offer every day. Conversely, do not keep a declining advert alive because its lifetime average remains attractive.
Record the highest spend level that still meets the payback rule. That is closer to current channel capacity than the audience size shown in the advertising interface.
Include inventory in the ad decision
Ads can create orders faster than stock can be replenished. Forecast sizes, variants and returns by acquired cohort. A campaign that sells only the most available size may look strong while leaving unbalanced inventory.
Set a stock floor that pauses or redirects spend before customer experience deteriorates. Include the deposit and lead time for the next production run in the payback window. If advertising cash returns after the supplier deposit is due, the business needs working capital even when lifetime contribution is positive.
Decide what counts as durable
Use a written rule such as:
- Three consecutive broad-audience cohorts reach positive contribution within 60 days.
- Return and complaint rates remain inside the product allowance.
- CAC stays below the limit after a 25% budget increase.
- No single creative supplies more than half of new-customer contribution.
- Stock can support the spend without an unfunded purchase order.
These thresholds are illustrative. Choose periods and concentration limits that fit your buying cycle. The discipline matters more than the exact number.
What to do over the next 60 days
This week, reconcile three months of advertising with settled orders. Split founder, retargeting, narrow and broad acquisition. Calculate contribution and returns by first-purchase cohort.
In week two, choose one weak cell in the matrix and one change. Cap spend at an amount you can lose without affecting supplier, tax or payroll commitments.
Review first-order results after the return window and repeat results at day 60. Increase spend only when three cohorts meet the rule. If the broad audience repeatedly fails, stop calling the launch spike scalable and repair the offer before buying more reach.
Related guides
Frequently asked questions
What return on ad spend should I target?
There is no universal target because gross margin, tax, returns, fulfilment and repeat buying differ. Calculate the revenue multiple that reaches your contribution requirement. If a £40 order leaves £16 before advertising, break-even first-order return on ad spend is £40 ÷ £16 = 2.5, before fixed overhead. If platform revenue includes VAT or later refunds, the required reported multiple is higher. Use the dashboard for diagnosis, then make budget decisions from settled order data. Label external benchmarks as indicative and verify them against your own product economics.
How long should I wait before judging a Facebook campaign?
Wait long enough to capture normal delivery, return and early repeat behaviour, while setting a spend cap in advance. First-order conversion may be visible within days, but retained contribution can take weeks. A product replenished monthly needs a longer cohort view than a one-off purchase. Do not leave a clearly uneconomic campaign running merely to satisfy an arbitrary learning period. Define the minimum sample and maximum loss together. Review tracking and stock problems immediately, then make the commercial decision when the relevant customer behaviour has occurred.
Should retargeting be included in customer acquisition cost?
Yes, but avoid crediting it with creating demand that prospecting, email or organic activity supplied. Calculate total new-customer advertising spend divided by genuinely new customers, then inspect prospecting and retargeting separately for diagnosis. Platform attribution can assign the same order to several touches. Use first-party order data and a consistent attribution rule. If retargeting is stopped, some customers may still buy, so its reported return is not all incremental. Run controlled holdouts where scale permits, or use conservative assumptions when it does not.
Can repeat purchases rescue a loss-making first order?
They can when observed cohorts repeat soon enough and with sufficient contribution. Do not rely on a projected lifetime value built from a few enthusiastic customers. Calculate cumulative contribution by cohort at 30, 60 and 90 days, using retained orders. Compare payback with stock and cash obligations. A six-month payback may be profitable in theory but unaffordable when suppliers need deposits every eight weeks. Set a maximum payback period from your cash cycle. If repeat behaviour worsens as audiences broaden, reduce the amount you are willing to lose initially.
How do I know whether an advert or the product caused returns?
Compare return reasons and rates by creative, audience, product and size. An advert promising opacity, fit or comfort may attract buyers whose expectations differ from the product delivered. The product page and size guidance can also create mismatch. Read customer comments and inspect returned items rather than relying on one reason code. Test corrected wording or guidance with the same product before changing manufacturing. If defects cluster by batch across acquisition sources, the product is the stronger suspect. Stop advertising a claim when it repeatedly creates disappointed buyers.
Should I stop ads when organic sales are growing?
Not automatically. Determine whether advertising creates incremental customers, supports search demand or mainly captures people already intending to buy. Reduce spend in controlled periods or regions and observe total new-customer contribution, not platform revenue alone. Organic growth can make blended acquisition economics stronger even when paid CAC rises, but paid reporting may overclaim it. Keep the channel when it adds profitable demand or valuable testing within your cash limit. Stop or narrow it when total contribution remains unchanged after spend, or when stock and service cannot support additional orders.
Comments (0)