The 365-day strategy and the hypothesis register: prioritization, a plan by area and defense before the owner and the bank
How after the first 72 hours to turn urgent measures into a plan for the year: build a hypothesis register with money, confidence, cost and hours, choose what to do for 30, 90 and 365 days, name owners and verification metrics and defend the plan before the owner, the bank or the board. The manager uses the document to decide which hypotheses to fund and which to postpone and how to check that the plan works.
Why this document
Crisis measures buy time but do not change the causes. Without a plan for the year the brand returns to the same state in a few months: discounts are deeper than plan again, stock accumulates again, suppliers get a prepayment again. The hypothesis register makes the plan verifiable: each assumption has money, probability, term, owner and a metric showing whether it worked. The document is read by the owner, the finance director, the heads of directions and the bank, which needs a schedule and grounds for trust.
At a glance
- What is this document and why is it needed?
- How after the first 72 hours to turn urgent measures into a plan for the year: build a hypothesis register with money, confidence, cost and hours, choose what to do for 30, 90 and 365 days, name owners and verification metrics and defend the plan before the owner, the bank or the board. The manager uses the document to decide which hypotheses to fund and which to postpone and how to check that the plan works.
- Who owns the document and how often is it rebuilt?
- Led by the finance director together with the heads of directions; the register is updated weekly, the plan is revised monthly and after each hypothesis test; handed to the owner and the bank at the end of the first week after the “72 hours” route and then monthly.
- What mistakes are the most common?
- Brand: hypotheses with an effect in half a year get into the plan → in the crisis months they do not help the money → structural hypotheses are moved to the 365-day horizon. Retailer: the register is filled with ideas without owners and test dates → after a month nobody knows what worked → the rule: no owner, metric and date, no hypothesis. Brand: confidence of 80-90% for all hypotheses → the order is set only by effect and hours → set confidence by a test, without a test not above 50%.
For whom
- Brand
- A brand builds the plan around the collection cycle: what to do with last season’s leftover, what to put into the purchase, which factories to keep, which channels to grow. The main thing for a brand: hypotheses about purchasing and production have a long time to effect, so they are decided in 30 and 90 days and not in 365.
- Retail
- A retailer builds the plan around the network and the assortment: which stores to grow and cut, what share of the assortment to hold with reliable suppliers, how to renegotiate rent. The main thing for a retailer: hypotheses about format and leases give effect slowly while price, promotions and stock management give effect fast.
What it consists of
Four horizons of the plan
- The first 72 hours: measures that cannot be postponed (money, obligations, contact with partners); their result is the register and the list of levers.
- 30 days: fast hypotheses with a money effect within a month: the price of hits, narrowing advertising, sale of dead stock, moving orders. This is the basis of the cash plan.
- 90 days: hypotheses with a medium term: a reorder of selling sizes, a revision of the platform terms, repair of cards, the first changes of the assortment.
- 365 days: structural hypotheses: a new channel, a new line, a change of factory, a reduction of space. Their effect is outside the crisis horizon but without them the brand will return to the same position.
- Brand: the first things recorded in the register are moving orders, a dead-stock sale, a second sewing source and a revision of the share of channels with commission.
- Retailer: the first things recorded in the register are weak stores, supplier and lease terms, promotions on slow stock and the delivery schedule of leaders.
Hypothesis register: template columns
- The base of the table: a number, an area (money, sales, stock, suppliers, team), a hypothesis in one phrase “if we do X, we get Y”, the mechanism (why it will work).
- Money and probability: the effect over the horizon in rubles, confidence in percent, one-off cost, team hours, days to effect. The expected net and the return per hour are calculated from them.
- Management: an owner, a verification metric (which number will show the result), a test date, a stop criterion, the actual result and the revised confidence.
- The rule: a hypothesis without an owner, a metric and a test date does not enter the plan. It protects against a wish list.
Plan by area
- Money: what we do with the reserve and credit relations, how we repay deferrals; the metric: the runway and the deviation of actual cash from plan by week.
- Sales: price, discounts, channels, hits; the metric: the share of sales at full price, sell-through, conversion, returns.
- Stock: what we sell, what we do not reorder, what safety stock the hits have; the metric: stock in weeks of sales, the share of dead stock.
- Suppliers: terms, a second source, the schedule; the metric: the share of orders on time, the unit price, dependence on one factory.
- Team: owners and motivation; the metric: hypotheses done on time and in budget.
Defending the plan before the owner and the bank
- The structure of the defense: the situation in numbers (cash, EBITDA, the loss map), causes, what was done in 72 hours, the plan by horizons, cash by week with the expected and the conservative paths, risks and what we do if a hypothesis fails.
- Brand: the bank and the investor are shown the schedule of deferral repayment, the dynamics of stock and the purchase plan; trust is built on the fact that last week’s plans were fulfilled.
- Retailer: the bank and landlords are shown turnover by store, the points’ contribution and the plan for the network; readiness to close loss-making points raises trust in the rest of the plan.
Where the data comes from
| Data source | Fields |
|---|---|
| Results of the first 72 hours | the reserve, the week of the break, the list of levers, the first version of the register, the loss map |
| Management report and plan versus actual | revenue, margin, costs, profit bridges, deviations by week |
| Direction teams | an estimate of effect, confidence, cost and hours for each hypothesis, owners, metrics and dates |
How to build it step by step
- Step 1: Move into the register everything left from the 72 hours: levers, ideas, the team’s remarks; give each a number and an area.
- Step 2: Brand: add hypotheses about purchasing, factories and the collection (a reorder, moving orders, a second source) with honest times to effect.
- Step 3: Retailer: add hypotheses about the network and suppliers (cutting points, rent, supplier terms, promotions) with honest times.
- Step 4: For each hypothesis fill in the money effect, confidence, cost, hours and days to effect; do not set confidence above 50% without a test.
- Step 5: Calculate the expectation (effect × confidence), the net after cost and the return per hour; drop what does not pay back or gives an effect after the horizon.
- Step 6: Build the plan in descending return while hours and budget last; leave the rest in the queue of the next round.
- Step 7: Brand: check the plan against the collection cycle and the factory terms: a purchase hypothesis that does not make the order window moves to the next season.
- Step 8: Retailer: check the plan against the season calendar and lease dates: a network hypothesis that does not make the termination date loses its meaning.
- Step 9: Assign owners, metrics and test dates; at the end of each week update the register: the fact, the revised confidence, the decision (continue, stop, scale).
- Step 10: Prepare the defense of the plan: the situation, causes, measures, the plan by horizons, cash by week, risks; rehearse the “what if it does not work” questions.
Formulas
| Measure | Formulas | Note |
|---|---|---|
| Expected net of a hypothesis | Net = effect × confidence − cost | If the net is not positive the hypothesis does not enter the plan; the effect is counted over the horizon, not over the year. |
| Return per hour | Return = net ÷ hours of work | Hours must be counted honestly, with approvals and waiting; there are no zero hours. |
| Plan load | Load = Σ plan hours ÷ team hours | Above 95% the plan risks not being fulfilled; leave a reserve. |
Formulas in notation
- Expected net and return
E = e·p − k, r = E / h
e: effect; p: confidence; k: cost; h: hours; E: expected net; r: return per hour.
Worked example
The brand’s hypothesis register after prioritization (the team gives 120 hours, a budget of RUB 800,000, a horizon of 90 days)
| Hypothesis | Effect, RUB | Confidence, % | Cost, RUB | Hours | Net, RUB | Return per hour, RUB | Status |
|---|---|---|---|---|---|---|---|
| Move part of the supplier orders to the next season | 2,000,000 | 65 | 0 | 14 | 1,300,000 | 92,857 | in the plan |
| Narrow advertising to channels where the ad share is not above the target | 1,800,000 | 70 | 0 | 16 | 1,260,000 | 78,750 | in the plan |
| Raise the price of the hits by 3% | 1,100,000 | 50 | 0 | 10 | 550,000 | 55,000 | in the plan |
| Sell dead stock through channels outside the main one | 2,400,000 | 60 | 150,000 | 30 | 1,290,000 | 43,000 | in the plan |
| Preorder for the loyal base with a discount | 900,000 | 55 | 80,000 | 20 | 415,000 | 20,750 | in the plan |
| Renegotiate commission and logistics with the platform | 1,500,000 | 30 | 0 | 24 | 450,000 | 18,750 | in the plan |
| Reorder fast styles in the selling sizes | 3,200,000 | 40 | 600,000 | 40 | 680,000 | 17,000 | no hours or budget left |
| Fix product cards and photos on the platform | 700,000 | 45 | 50,000 | 36 | 265,000 | 7,361 | no hours or budget left |
| Reduce the rented space | 1,300,000 | 25 | 400,000 | 60 | −75,000 | −1,250 | does not pay back |
| A new product line | 6,000,000 | 20 | 1,500,000 | 160 | −300,000 | −1,875 | effect after the horizon |
Six hypotheses made the plan: a net of RUB 5.27 million at 114 hours and a cost of RUB 230,000; RUB 1.81 million will arrive in the first 72 hours. The reorder of fast styles and the platform cards did not fit by hours and budget, the “new line” is cut by timing, the “reduction of space” does not pay back. The numbers are illustrative and calculated by code (the hypoprio calculation).
Figures
Data as a table
| Measure | Return per hour |
|---|---|
| Move part of the supplier orders… | 92.9 |
| Narrow advertising to channels w… | 78.8 |
| Raise the price of the hits by 3% | 55 |
| Sell dead stock through channels… | 43 |
| Preorder for the loyal base with… | 20.8 |
| Renegotiate commission and logis… | 18.8 |
| Reorder fast styles in the selli… | 17 |
| Fix product cards and photos on … | 7.4 |
| Reduce the rented space | −1.2 |
| A new product line | −1.9 |
Return per hour by hypothesis, RUB thousand per hour. First come the hypotheses that give more money per hour of work: moving orders and narrowing advertising; large but long hypotheses yield.
The year by horizons: when what is tested. Fast hypotheses are tested in the first weeks, medium ones by the end of the quarter, structural ones during the year; checkpoints every week.
| Stage | Start, wk | Length, wk |
|---|---|---|
| 72 hours: the route | 0 | 1 |
| 30 days: fast hypotheses | 1 | 4 |
| 90 days: medium hypotheses | 4 | 9 |
| 365 days: structural hypotheses | 13 | 39 |
| Plan revision and defense | 4 | 48 |
The register as a cycle: hypothesis, test, decision. Each hypothesis goes through the cycle; the result changes the confidence of the remaining hypotheses.
- Hypothesis and estimate → Prioritization
- Prioritization → Test and metric
- Test and metric → Decision: continue, stop, scale
- Decision: continue, stop, scale → Update confidence
- Update confidence → Prioritization
Specifics by category
- Apparel
- Apparel: hypotheses about stock and price give a fast effect; structural ones concern the collection calendar and the share of basic styles; seasonality forces the term to be tied to the order window. For seasonal styles plan two checkpoints: the middle of the season and the start of the clearance so that the decision on leftovers is not late.
- Footwear
- Footwear: most hypotheses on purchasing and the size run have a long time to effect; price and sale go into the 30- and 90-day plan while changes of lasts and factory go into the 365 days.
- Accessories
- Accessories: hypotheses on the assortment and channels are tested faster; wholesale and gift sales allow testing demand with a small lot before a large purchase. Small lots and fast tests allow putting hypotheses into the 30-day horizon and scaling winners quickly without risking the stock of a whole season.
- Kids
- Children’s: hypotheses are tied to the school season and children’s growth; an error in sizes cannot be corrected until the next season, so confidence in purchasing without a test is kept low. The school and growth calendar sets hard dates so assign the owner and the test date in advance and do not move them without a reason.
- Men
- Men’s: steady demand allows planning hypotheses on the base range and price over a long horizon; the effect is more often in margin and purchasing than in fashion. So the 365-day plan for men’s is built around the base and prices: fewer hypotheses about fashion, more about purchasing, margin and supplier terms.
- Women
- Women’s: a high share of fashion makes assortment hypotheses less certain, so a reorder of selling sizes and preorders are tested with small lots. For women’s put more fast tests with a stop criterion: fashion changes quickly and waiting for the quarter’s result means missing the window.
Common mistakes
- Brand: hypotheses with an effect in half a year get into the plan → in the crisis months they do not help the money → structural hypotheses are moved to the 365-day horizon.
- Retailer: the register is filled with ideas without owners and test dates → after a month nobody knows what worked → the rule: no owner, metric and date, no hypothesis.
- Brand: confidence of 80-90% for all hypotheses → the order is set only by effect and hours → set confidence by a test, without a test not above 50%.
- Retailer: the plan is not revised after hypothesis tests → resources go into failed ideas → update the fact and the decision weekly.
- The defense of the plan is built on the expected path → when one lever fails the bank loses trust → show the conservative path too and the actions on failure.
Russian specifics
- Brand: agree the plan figures and the schedule for the bank with the terms of the credit contract (covenants, the reporting procedure); in Lab thresholds are not given as fact, check the contract.
- Retailer: lease changes, closing points and reductions are regulated by the contract and the law; before putting such hypotheses into the plan agree the terms and compensation with a lawyer.
- Brand: accounting statements and provisions (inventory at the lower of the values, provisions for the fall in value) in a dead-stock sale touch FSBU 5/2019; formalize write-off decisions with an accountant. Rule: Order 180n of 15.11.2019: Ministry of Finance Order of 15.11.2019 No. 180n (FSBU 5/2019 Inventories) (rule check not completed).
- Retailer: the tax consequences of closing points, returning rent and writing off goods are checked with a tax adviser before the decision.