Mathematically, neither lever has an inherent advantage. A 10% increase in average order value produces the same revenue uplift as a 10% increase in purchase frequency, provided customer numbers and the other inputs remain unchanged.
Commercially, however, the two strategies can produce very different outcomes.
Larger orders may require discounts, premium offers, or changes in customer mix. More frequent purchases may require lifecycle programs, additional fulfillment capacity, or a product customers genuinely need more often.
The deciding factor is contribution: which achievable change produces more value after its costs and side effects?
A sensitivity model helps compare the possibilities. An evidence-based experiment determines which possibility deserves investment.
Start With the Revenue Equation
For a defined customer population and period:
Revenue = Customers × Purchases per customer × Average order valueAverage order value, or AOV, measures revenue per transaction. Purchase frequency measures the average number of orders each customer places during a specified period. Both metrics should use compatible data and timeframes.
AOV = Revenue / Orders
Purchase frequency = Orders / CustomersFor historical analysis, these definitions reconcile revenue directly. For forecasting, they become assumptions about future customer behavior.
That distinction matters: an equation can show what an improvement would produce without establishing that the improvement is achievable.
A Shared Baseline for Comparing the Two Levers
Consider an illustrative six-month customer cohort.
| Baseline input | Assumption |
|---|---|
| Customers | 400 |
| Average order value | $1,250 |
| Purchases per customer within six months | 2 |
| Total orders | 800 |
| Revenue | $1,000,000 |
The calculation is:
400 × 2 × 1,250 = 1,000,000For the initial comparison, hold customer count constant. Also assume the purchase-frequency estimate already accounts for when customers enter the cohort and how much time they have to purchase.
Scenario 1: Increase Average Order Value
Suppose AOV increases by 20%, from $1,250 to $1,500.
With 400 customers and two purchases per customer:
400 × 2 × 1,500 = 1,200,000Revenue increases by $200,000.
Possible initiatives include:
- Offering a higher-value package.
- Introducing relevant add-ons.
- Creating bundles.
- Improving premium-tier positioning.
- Changing pricing.
- Targeting customers with larger requirements.
These initiatives are not economically equivalent.
A higher price may increase margin but affect conversion. A discounted bundle may increase revenue per transaction while reducing margin. A shift toward larger customers may require a different sales process.
The forecast needs to model the mechanism behind the higher AOV—not simply replace one number with another.
Scenario 2: Increase Purchase Frequency
Now hold AOV at $1,250 and increase average purchases per customer by 20%, from two to 2.4.
400 × 2.4 × 1,250 = 1,200,000Revenue again increases by $200,000.
An average of 2.4 purchases does not mean an individual customer makes a fractional purchase. It describes the cohort: 400 customers generate 960 orders rather than 800.
Possible initiatives include:
- Timely replenishment reminders.
- Customer education.
- Relevant cross-selling.
- Lifecycle communications.
- Reactivation programs.
- Recurring purchase options.
- Improvements that make repeat buying easier.
The key question is whether these initiatives enable valuable additional purchases—or merely encourage customers to split, accelerate, or discount purchases they would have made anyway.
Revenue Sensitivity Table
The following table holds the customer population at 400 and varies AOV and purchase frequency.
| Purchases per customer over six months | AOV: $1,000 | AOV: $1,250 | AOV: $1,500 | AOV: $1,750 |
|---|---|---|---|---|
| 1.5 | $600,000 | $750,000 | $900,000 | $1,050,000 |
| 2.0 | $800,000 | $1,000,000 | $1,200,000 | $1,400,000 |
| 2.4 | $960,000 | $1,200,000 | $1,440,000 | $1,680,000 |
| 3.0 | $1,200,000 | $1,500,000 | $1,800,000 | $2,100,000 |
All figures are illustrative scenarios, not forecasts. Customer count remains unchanged, and all modeled purchases occur within the six-month period.
The table highlights three relationships.
First, different combinations can produce the same revenue. At 400 customers, $1.2 million can come from:
- Two purchases at $1,500.
- 2.4 purchases at $1,250.
- Three purchases at $1,000.
Second, equal percentage improvements in either lever produce equal revenue improvements when everything else is fixed.
Third, changing both levers creates a multiplicative effect.
If AOV and frequency each increase by 20%:
1.2 × 1.2 = 1.44Revenue increases by 44%, not 40%.
That combined result is valid only if the two improvements can coexist.
The Levers Can Work Against Each Other
AOV and purchase frequency are connected behaviors, not independent controls.
Larger orders may reduce purchase frequency
Suppose a bundle raises AOV from $1,250 to $1,500, but customers need to reorder less often. Frequency falls from two purchases to 1.6.
400 × 1.6 × 1,500 = 960,000AOV rises by 20%, but revenue falls by 4%.
The bundle may still improve fulfillment efficiency or customer satisfaction. However, the higher AOV alone would give an incomplete picture.
More orders may reduce average order value
Suppose frequency rises from two to 2.5 purchases, while AOV falls from $1,250 to $1,000.
400 × 2.5 × 1,000 = 1,000,000Customers place 25% more orders, but revenue does not change.
If each order creates additional handling, payment, or support costs, the business may be worse off despite the stronger frequency metric.
Evaluate revenue per customer alongside both metrics:
Period revenue per customer = AOV × Purchase frequencyCompare the Economics Behind Equal Revenue Outcomes
Even when both strategies generate $1.2 million, their financial results may differ.
Consider two illustrative programs compared with the $1 million baseline:
| Measure | Baseline | Higher-AOV program | Higher-frequency program |
|---|---|---|---|
| Customers | 400 | 400 | 400 |
| Average order value | $1,250 | $1,500 | $1,250 |
| Purchases per customer | 2 | 2 | 2.4 |
| Revenue | $1,000,00 0 | $1,200,000 | $1,200,000 |
| Gross margin | 45% | 42% | 45% |
| Gross profit | $450,000 | $504,000 | $540,000 |
| Additional program costs | — | $25,000 | $40,000 |
| Incremental contribution versus baseline | — | $29,000 | $50,000 |
Program costs are additional to baseline operations and excluded from gross profit. Assume all relevant delivery costs are reflected in the stated margins.
In this scenario, the higher-AOV program uses an offer mix that reduces gross margin. Its revenue uplift is identical to the frequency program’s, but its incremental contribution is lower.
The result could reverse under different assumptions. A price increase that preserves demand might improve margin. A repeat-purchase program dependent on expensive incentives might reduce it.
The comparison should use:
Incremental contribution = Scenario gross profit − Baseline gross profit − Additional program costsDo not apply the original margin to the revenue uplift without checking whether the initiative changes delivery costs, discounts, or product mix.
Respect the Planning Period
Purchase frequency has no useful meaning without a timeframe.
Two purchases per month, two purchases in six months, and two purchases over a customer’s lifetime describe different economics.
This is especially important for newly acquired customers. A customer acquired in month six has less time to reorder than a customer acquired in month one.
Use acquisition cohorts to assess repeat purchases and spending over comparable periods. Shopify’s customer analysis guidance similarly recommends grouping customers by their first purchase and reviewing subsequent purchasing behavior.
Also distinguish additional demand from timing shifts.
A campaign may bring a purchase forward from next quarter into the current quarter. That improves the current-period result, but it does not necessarily increase revenue over a longer horizon.
Choose the observation window accordingly.
Which Lever Deserves the Next Experiment?
Use the following framework to identify the more credible opportunity.
| Decision factor | Favor an AOV experiment when… | Favor a frequency experiment when… |
|---|---|---|
| Customer need | Customers could benefit from a larger or more valuable solution | Customers have recurring or adjacent needs |
| Evidence | Research or transaction data suggests relevant upgrade potential | Cohort data shows repeat-purchase opportunities |
| Main friction | Packaging or offer structure limits order value | Timing, awareness, or buying friction limits reordering |
| Economics | The larger order preserves sufficient contribution | Additional orders remain attractive after fulfillment and incentives |
| Timing | The offer can affect purchases within the target period | Customers have enough time and reason to purchase again |
| Primary risk | Lower conversion or reduced future purchasing | Smaller orders, discount dependence, or purchase pull-forward |
| Key guardrails | Conversion, margin, refunds, subsequent spending | AOV, contribution per customer, order costs, subsequent-period revenue |
Neither column establishes a universal winner. The better experiment addresses a documented opportunity with acceptable cost and measurable downside.
Questions to Ask Before Approving Either Scenario
For a higher-AOV plan
- Why would customers choose the larger purchase?
- Does the increase come from price, quantity, product mix, or customer mix?
- Could conversion decline?
- Will the offer reduce future orders?
- What happens to gross profit per order and per customer?
- Are implementation and sales costs included?
For a higher-frequency plan
- Does the customer genuinely need to buy more often?
- Which customers are eligible for another purchase?
- Can the additional purchases occur before the deadline?
- Are incentives creating value or subsidizing existing behavior?
- Will smaller orders increase operating costs?
- Does the effect persist beyond the campaign?
For either experiment, define success using period contribution per eligible customer—not just the metric being optimized. Keep the eligible customer population visible so that declining conversion or lost customers do not disappear from the comparison.
Use the Forecast to Set a Testable Requirement
A sensitivity table helps leadership understand what different combinations would produce.
It does not establish which combination customers will accept.
A practical process is to:
- Define the customer cohort and period.
- Establish baseline AOV and purchase frequency.
- Model plausible changes in each.
- Include interactions, margin changes, and program costs.
- Identify the minimum improvement required to justify investment.
- Test the proposed mechanism.
- Update the forecast using observed customer behavior.
In a scenario tool such as PlotGTM, AOV and purchase-frequency controls can make these trade-offs visible. The important step is attaching evidence and operational conditions to every proposed change.
The strongest growth lever is not inherently a larger order or a more frequent purchase. It is the customer behavior the business can improve credibly, profitably, and within the period that matters.