Data & Analytics

How to use simple cohort-driven clv forecasting to set reward ceilings for seasonal promotions

How to use simple cohort-driven clv forecasting to set reward ceilings for seasonal promotions

Seasonal promotions are a huge opportunity — and a huge risk. Too generous, and you eat margin without improving long-term value. Too stingy, and you miss sales and the chance to deepen customer relationships. Over the years I’ve found one pragmatic way to navigate that trade-off: a simple cohort-driven CLV forecasting approach that gives you an evidence-based ceiling on how much you can afford to spend on rewards or acquisition for a given promo.

Why cohort-driven CLV beats one-size-fits-all rules

Many SMEs default to blanket rules such as “give 10% off” or “offer free shipping over £50” because they’re easy to implement. But customers aren’t all the same: first-time buyers behave differently from repeat purchasers, and seasonal cohorts (e.g. holiday shoppers vs. summer buyers) can show distinct retention and spend patterns.

Cohort-driven CLV (Customer Lifetime Value) forecasting splits your customer base into groups that share a common starting point — typically their acquisition month or campaign — and projects value for each group separately. The advantage is clear: you can set promo ceilings based on realistic future revenue for the specific customers you’ll win, rather than a vague average that might under- or over-estimate value.

What I use in practice: the simple cohort CLV model

Here’s a lightweight model I use with clients when they need a rapid yet defensible estimate. It requires only basic data and gives an actionable reward ceiling.

Inputs I look for:

  • Acquisition cohort (e.g. month customers first bought)
  • Average order value (AOV) for the cohort
  • Repeat purchase rate per period (typically monthly or quarterly)
  • Gross margin (%) on product sold
  • Discount or promo mechanics under consideration
  • The basic steps:

  • Build cohort retention curve: calculate the proportion of the cohort that makes a purchase in each subsequent period.
  • Project future revenue by multiplying the period retention by AOV, summing across a reasonable horizon (6–24 months depending on your category).
  • Apply gross margin to get projected gross profit per cohort.
  • Subtract customer acquisition or promo cost to test affordability — the leftover gross profit is what you’ve effectively paid for future value.
  • Formula translated into a simple spreadsheet:

    For each cohort month, do the following in separate columns:

  • Period revenue = Cohort size × Retention rate(period) × AOV
  • Projected revenue (N periods) = Sum(period revenue t=0..N)
  • Projected gross profit = Projected revenue × Gross margin
  • Reward ceiling per acquired customer = Projected gross profit – required minimum profit (or target ROAS threshold)
  • Example values I commonly use:

    Cohort size 1,000 customers
    Initial AOV £60
    Gross margin 45%
    Retention (month 1,2,3...) 40%, 20%, 12%, 8%, 6% ...

    Using those numbers and projecting 12 months, you get a sum of expected revenue per acquired customer. Multiply by margin to get gross profit per customer, then decide how much of that you can allocate to rewards.

    Practical tips for setting the reward ceiling

    When I build ceilings with teams I always apply a few pragmatic constraints:

  • Use a conservative projection horizon. If your category sees long gaps between purchases (e.g. furniture, travel), extend the horizon. For fast-moving FMCG or fashion, 6–12 months is usually enough.
  • Prefer cohort-specific AOV and retention. Holiday cohorts often have higher AOV but lower retention. Use the cohort’s own metrics, not site-wide averages.
  • Account for cannibalisation. Some reward-induced spend replaces future full-price spend. I discount projected uplift by an estimated cannibalisation rate — I often start with 20–30% unless I have test data.
  • Include operational and fulfilment costs. A free gift mechanically has a product cost and handling cost — include those in the ceiling calculation.
  • How to handle acquisition vs. retention promos

    The cohort CLV approach works for both, but the math and acceptable ceilings differ.

  • Acquisition promos (e.g. paid ads with a first-order discount): Base the ceiling on projected CLV of new customers in the specific acquisition channel. If Facebook-acquired customers have lower retention, give them a lower maximum reward or use a post-purchase incentive (e.g. discount on next purchase) to improve retention.
  • Retention promos (e.g. email-only seasonal sale for existing customers): Use observed repeat behaviour of existing cohorts. Often you can be more generous because you’re increasing frequency among already valuable customers.
  • Testing and iterating: the A/B playbook I recommend

    Numbers are helpful, but nothing replaces an experiment.

  • Run an A/B test on the promotion: control (no change) vs. promo A (smaller reward) vs. promo B (larger reward at ceiling). Track incremental revenue, repeat purchases and margin impact.
  • Measure 30/90/180 day outcomes depending on your purchase cadence. Don’t over-interpret immediate uplift — the real payoff is in repeat behaviour.
  • Use cohort analysis for both test and control groups. This ensures you’re comparing like-for-like customers acquired in the same period.
  • Example: a seasonal promo for a mid-sized DTC brand

    One client — a DTC homewares brand — wanted to run a winter promotion. Their data showed:

  • Black Friday cohort AOV = £80, month 0 retention = 35%, month 1 = 18%, month 2 = 10%, and long tail.
  • Gross margin ~50%.
  • Projected 12-month gross profit per new customer came to roughly £28. If the team wanted at least £8 net contribution per customer after the promo, that left ~£20 to fund acquisition and rewards. They split that into £12 maximum paid CAC and £8 maximum reward value per customer.

    They then ran two promo arms: a straight £8 off first order vs. free £8 gift with purchase that cost them £5 after fulfilment. The gift performed better for retention (higher repeat rate) and delivered stronger long-term CLV — proving the cohort forecast and the operational costs gave a useful ceiling to test.

    Common pitfalls to avoid

  • Using a single average CLV for all customer types.
  • Ignoring variable costs (fulfilment, returns, customer service load) tied to promotions.
  • Not testing — assuming the ceiling is the final word rather than a hypothesis to validate.
  • If you want, I can sketch a ready-to-use spreadsheet template with cohort columns and the formulas shown above so you can drop in your numbers and get reward ceilings for any seasonal plan. I often share that with clients to speed up decision-making and to move from opinion to evidence within a few hours.

    You should also check the following news:

    How to run a 30-day retention lift test using till data and one email sequence
    Customer Retention

    How to run a 30-day retention lift test using till data and one email sequence

    I run a lot of experiments for small retailers and hospitality brands, and one of the simplest...

    Jul 08 Read more...
    Which microreward timing lifts second purchase fastest: day‑0 free gift, day‑7 reminder credit or triggered sms
    Customer Retention

    Which microreward timing lifts second purchase fastest: day‑0 free gift, day‑7 reminder credit or triggered sms

    I recently ran a small-but-rigorous test to answer a question I get asked all the time: which...

    Jun 21 Read more...