Separate recurring seasonality from trends and one-off spikes by comparing buyer calendars, search and enquiry patterns, competitor activity and low-period cash.
Short answer: Look for the same rise and fall at the same point in at least two annual cycles across three independent signals: buyer deadlines, search or enquiry behaviour, and competitor capacity or pricing. If 24 months of comparable evidence does not exist, treat seasonality as unproven, test through one complete buying cycle and fund the lowest plausible 13 weeks. Do not launch from an annual average or one peak month.
A demand spike can come from weather, promotion, a news event, supply failure or genuine seasonality. Only seasonality repeats because a calendar, climate or institutional cycle repeatedly changes customer behaviour.
The distinction matters because a seasonal business can be healthy while losing cash for several months. A temporary spike can persuade you to add fixed costs just as demand returns to normal.
Use the Seasonality Triangulation Method
The Seasonality Triangulation Method requires three different views of the same pattern. No single data series is conclusive.
- Buyer calendar: What to collect: Deadlines, events, budgets, maintenance or replacement moments; Strong seasonal pattern: The same trigger recurs at a predictable point
- Search and enquiries: What to collect: Weekly queries, calls, quotes or bookings by source; Strong seasonal pattern: Similar rise and fall repeats after adjusting for promotion
- Competitor behaviour: What to collect: Lead times, opening, staffing, stock and price changes; Strong seasonal pattern: Capacity tightens and loosens around the same period
Two years lets you observe recurrence once. More cycles strengthen confidence, especially where weather varies. If only one year exists, label the conclusion provisional and plan from the low case.
Separate seasonality from other patterns
Classify the shape before forecasting it.
- Seasonality: What it looks like: Repeats at similar calendar points; Common cause: Weather, holidays, budgets or annual routines; Planning response: Plan peak capacity and trough cash
- Trend: What it looks like: Moves mainly in one direction; Common cause: Structural adoption or decline; Planning response: Test whether the mechanism persists
- Event spike: What it looks like: Sharp temporary rise; Common cause: News, disruption or one-off occasion; Planning response: Avoid permanent cost based on it
- Promotion effect: What it looks like: Rises when you spend or discount; Common cause: Acquired attention; Planning response: Include acquisition cost in demand
- Noise: What it looks like: Irregular movement without a stable mechanism; Common cause: Small samples or random variation; Planning response: Collect more evidence and stay flexible
My view is that a pattern without an identifiable buyer mechanism should not be called seasonal. A chart can repeat by coincidence. The customer calendar explains why it might repeat again.
Start with the buying trigger
Ask customers when they last bought, what happened beforehand and what date constrained the decision. A summer product may be ordered in winter by trade buyers. A January service may be researched in November.
Map four dates:
- Problem becomes noticeable
- Customer begins research
- Customer commits or pays
- Delivery occurs
Your marketing, cash and capacity seasons may therefore differ. Ordering stock for the delivery peak can be too late if supplier lead times begin months earlier.
Avoid asking “Is this seasonal?” in general. Reconstruct the last two purchases and compare with operational records where available.
Normalise what you can control
Enquiries fall when you stop answering, reduce promotion or run out of capacity. Competitor bookings may look seasonal because the provider closes for a month. Separate demand from availability.
For each period, record:
- Days open and delivery capacity
- Price and any discount
- Customer-acquisition spending and activity
- Stock or appointment availability
- Weather or event conditions relevant to the offer
Compare enquiries or sales per open day and by source. Do not interpret zero sales during zero availability as low demand.
Search estimates should be treated as indicative. Sources group queries and average periods differently. Use them alongside actual buyer and competitor evidence rather than as a precise forecast.
Calculate the trough before enjoying the peak
Use contribution, not revenue. A peak month can bring overtime, express supplies and higher acquisition cost that reduce the cash available for quiet months.
Calculate: Monthly operating result = paid transactions × contribution per transaction minus fixed business cash and required owner pay
Run the formula for each month or week. The lowest cumulative balance determines the seasonal reserve, not the average annual profit.
Some practitioners advise creating an off-season offer. That can help when the same buyer has a related need and you can serve it competently. It can also create a second unproven business. My view is to price and fund the core season first. Test any complementary offer on its own demand evidence.
Worked example: Theo's garden-furniture restoration service
Theo tests a garden-furniture restoration service while keeping costs flexible. During a 12-week spring period, he completes 36 paid jobs, or three a week. In a comparable 12-week autumn period, he completes 12, or one a week.
Average price is £360. Materials, travel, payment and sale-specific costs are £85, leaving contribution of: £360 minus £85 = £275 per job.
Using 4.3 weeks as an indicative month, spring contribution is: 3 jobs × 4.3 × £275 = £3,547.50.
Autumn contribution is: 1 job × 4.3 × £275 = £1,182.50.
Theo's fixed business cash plus required owner pay is £2,200 a month. Spring produces £3,547.50 minus £2,200 = £1,347.50 surplus. Autumn produces £1,182.50 minus £2,200 = negative £1,017.50.
If four low months resemble autumn, the cash requirement is: 4 × £1,017.50 = £4,070.
Theo cannot call one spring and one autumn a proven annual pattern. He checks buyer timing, comparable search and enquiry history, and competitor lead times for recurrence. Until a second cycle confirms it, he should preserve at least the £4,070 calculated trough plus a separate correction allowance and avoid fixed commitments based on spring alone.
His capacity is four jobs a week, so spring uses 75 per cent. A hotter peak might create more demand than he can fulfil, making advance booking and price more useful than extra fixed equipment.
Look for leading indicators
Bookings, quote requests, supplier orders and relevant searches can move before revenue. Identify which signal leads paid work consistently and by how many weeks.
A leading indicator lets you adjust temporary capacity and cash earlier. It is not a guarantee. Compare predicted bookings with completed, paid jobs and record cancellations.
Competitor lead times can also help. Longer waits may show demand, reduced capacity or both. Ask customers when they first contacted alternatives and why they delayed before interpreting the signal.
Decide when to launch
Launching just before a peak can produce fast sales and operational mistakes at once. Launching in the trough provides time to learn but may give a falsely pessimistic demand result.
Choose a test window that includes the lead-in and part of the peak, while keeping commitments reversible. Set price, response time and capacity before demand rises. Do not discount the peak merely to prove volume.
If premises, staff, stock or finance continue through the trough, calculate them for the full year. Employment, lease, tax, consumer and sector obligations vary by country and activity. Use qualified local professionals for specific commitments.
Related guides
Complete the seasonality check in 30 days
Collect at least 24 months of weekly or monthly proxy evidence where it exists, recording the source and changes in availability or promotion. Interview customers about their last two buying cycles and map the four trigger dates.
Then act in this order:
- Classify the pattern as seasonality, trend, spike, promotion or noise.
- Require an explanatory buyer mechanism and three supporting signals.
- Calculate contribution and cash for peak, normal and low periods.
- Fund the lowest plausible 13 weeks before adding fixed commitments.
- Run a paid test through a complete buying cycle and update the pattern.
Plan from the trough. Let the peak provide upside rather than rescue.
Frequently asked questions
How much historical data do I need to identify seasonality?
Use at least 24 comparable months where available because that shows whether a pattern repeats once. Three or more annual cycles are stronger when weather, events or small samples create volatility. Data quantity does not replace mechanism: link the rise and fall to a recurring customer trigger.
If no history exists, use related buyer records, search and competitor evidence, then treat the conclusion as provisional. The exception is a fixed calendar event with an obvious annual date, but the size and timing of demand can still change, so do not assume last year's volume repeats.
Is a seasonal business a bad business?
No. It can be attractive when peak contribution funds the trough, costs flex with demand and capacity can handle the busy period without harming quality. Calculate cash by week or month, not annual average, and preserve the maximum cumulative deficit.
Seasonality becomes dangerous when rent, finance or payroll continue while sales disappear, or when one bad peak ruins the year. The exception is a business intentionally operated for only part of the year with minimal off-season commitments. It still needs personal income and tax planning appropriate to the owner's circumstances.
How can I forecast seasonality with no sales history?
Triangulate buyer calendars, search or enquiry proxies and competitor behaviour, then run a reversible paid test through one complete buying cycle. Ask customers about actual previous purchases and the trigger dates. Use ranges and label assumptions. Do not copy another firm's monthly percentages without checking geography, price and capacity.
Your first cash plan should use the low case and preserve optional costs. The exception is taking over a verified existing business where detailed records transfer, but customer retention and your own operating changes can still alter the pattern.
How do I tell a trend from a seasonal peak?
A seasonal peak falls and returns near the same calendar point because a recurring mechanism drives it. A trend moves mainly in one direction across seasons. Compare like months across multiple years rather than consecutive months. Remove changes in price, promotion, availability and source where possible.
Ask what changed in customer behaviour. One year cannot distinguish them confidently. The exception is a clear external change, such as a new requirement, which may create a structural step. Verify its scope and duration with current authoritative information before treating it as permanent demand.
Should I launch during the busiest season?
Only if you can deliver reliably and have tested the core process before the peak. Busy periods provide fast evidence but can hide inefficiency because customers tolerate delays or founders work unsustainable hours. A short pre-peak test lets you correct scope, suppliers and response times. Set capacity and price before launch. The exception is an annual event that provides the only credible test window. Keep the offer small, cap orders and preserve refund or correction cash rather than accepting every sale to maximise first-year revenue.
Can I raise prices in the peak and lower them in the trough?
You can vary price when terms are clear, lawful and supported by customer value and capacity. Peak pricing may ration scarce slots and fund quiet months. Trough discounts can attract price-sensitive demand but may train customers to wait and reduce contribution when cash is already weak.
Test total contribution, not sales volume, and avoid unfair or misleading practices. The exception is contracted or regulated pricing, where changes may be restricted. Consumer, competition and sector rules vary by jurisdiction, so obtain qualified local advice before adopting material differential pricing.
Comments (0)