Ansoff matrix
Separate growth options by existing or new products and markets, then examine their demand and capability assumptions.
Use the matrix when growth ideas mix product and market changes. Define newness against one business and baseline, then distinguish penetration, product development, market development and diversification.
Current product and market boundaries, a baseline date, options, demand evidence and capability or resource constraints.
Four types of option with classification reasons and explicit evidence needs.
On this page
How to use it
- Define the business, baseline and existing/new product and market rules.
- Identify product and customer or geographic changes in each option.
- Classify consistently or leave an option unresolved; empty cells are acceptable.
- Check demand, capabilities, resources and dependencies for each option.
- Choose an evidence-gathering action and retain actual approval status.
Keep its limits in mind
A channel change is not necessarily a new market, nor a small feature change a new product. Quadrants are not risk probabilities, a fixed difficulty ranking or an investment recommendation. Existing combinations can fail. Keep empty or unclassified options rather than inventing plans; an option is not approval or achieved growth.
Sorting four options without changing the definition of a market
Fictional teaching example: Alder Lunch is reviewing its direction. S3 contains three provider cards (E1, Oct 5–9, 2026), Alder's 100-order ledger (E2), ratings by three fictional respondents (E3), capability records (E4), and supply/entry questions (E5). These are synthetic records, not market research. Proposed actions have not been carried out.
On a small screen, scroll across the diagram to read it.
Fix the baseline: P0 and M0 describe current lunches and individual Harbor office workers. P1 is a new single-person afternoon snack box; M1 is a new geography and student group. A web-to-app change alone does not meet this market-change rule.
Classify consistently: Reminders for current lunches fit penetration; snack boxes for current workers fit product development; lunches for Northbank students fit market development; snack boxes for those students fit diversification.
Result: Four classified options remain proposals. Demand and operational resources have not been established, no launch has occurred, and the quadrant supplies neither a success probability nor a funding priority.
P0/P1 and M0/M1 use one baseline. All four directions are unapproved, unlaunched options without risk probabilities or an automatic ranking.
Next step: The strategy lead should confirm product, customer and geography definitions, then examine the customer task and resources for one option before choosing a trial.
Three reasons to leave an option unclassified
First, stop when “new” changes meaning between options. Define the business and baseline date before discussing growth. A new ordering channel may serve the same people with the same lunch. A genuinely different customer task or offering needs its own explanation. Classification should follow a declared rule, not make a preferred proposal appear less demanding.
Second, do not force every idea into a populated quadrant. An unclear proposal may need clarification, and an empty cell may simply mean no credible option has been identified. The matrix organizes possibilities; it does not require four launches or establish that the business should diversify.
Third, pause if the quadrant is being used as an investment verdict. Familiar products and customers can still disappoint. Newness identifies questions about demand, capabilities, access, and resources; it does not supply success probabilities or a fixed order for funding. Keep approval separate from the discussion.
Distinguish growth direction from customers, capabilities, and economics
Use the matrix to identify what changes in an option. Follow with a different method if the missing evidence concerns whom to serve, how to deliver, or whether an operating model is coherent. None of these worksheets supplies demand evidence merely by being completed.
| Question and available input | Method | Output and limit |
|---|---|---|
| Several options mix product and market changes | Ansoff matrix | Consistent growth categories and evidence needs; no ranking. |
| Customer groups and needs require investigation | STP | Segmentation and target hypotheses; research remains necessary. |
| An option depends on an organizational capability | VRIO analysis | Resource conditions and gaps; no automatic market demand. |
| Delivery, partners and economics need to fit | Business Model Canvas | Connected operating assumptions; not a validated business. |
Classify Alder’s four options against one baseline
The complete fictional baseline is October 9, 2026: P0 is a single weekday lunch and M0 is individual Harbor office workers. P1 is a single-person afternoon snack box; M1 is Northbank college students, a change in both geography and customer group. The options are reorder reminders for P0/M0, snack boxes for P1/M0, lunches for P0/M1, and snack boxes for P1/M1. They are unapproved, unlaunched proposals. Demand, people, and budget are unknown.
Under these definitions, reminders for existing lunches fit penetration even if the message uses a different channel. Snack boxes for the same workers change the product. The same lunch offered to Northbank students changes the market. Snack boxes for those students change both. Calling all four “digital growth” would hide these different assumptions.
The table retains one investigation question per option. These are proposed checks, not reports of customer interviews or permission. The same definition must remain in use when an option is revised; if the product or market boundary changes, preserve the reason and reclassify rather than quietly moving it to a preferred cell.
| Option and classification | What changes under the baseline | Proposed evidence check |
|---|---|---|
| P0/M0 reminders: penetration | Neither product nor market | Investigate current reorder task and permission for reminders; no uplift known. |
| P1/M0 snack box: product development | Product only | Check whether workers need an afternoon offering and whether it can be supplied. |
| P0/M1 lunch: market development | Customer group and geography | Check student task, access and delivery conditions; existing capability may not transfer. |
| P1/M1 snack box: diversification | Both product and market | Investigate both task and supply assumptions; no quadrant probability. |
Review the option, the evidence, and the actual decision
The strategy lead should first confirm that the baseline describes the current business accurately. Existing order records can help describe current activity, but they do not validate snack-box demand or represent Northbank students. Keep the different populations and offerings identifiable when collecting further material.
Before any trial, name the question, intended participants, access requirements, operational constraints, and decision-maker. A conversation about a snack box can clarify a task without demonstrating a purchase. A request to contact students does not establish campus access, permission, staffing, or a funded launch.
At review, retain what was actually learned, which assumptions remain, and whether the option itself changed. An option may need to be abandoned, narrowed, or investigated further. Record any authorized next step with its scope and conditions. No completed research, investment decision, sales increase, or successful entry is supplied in this example.
If a later trial proceeds, distinguish invitation, interest, order, fulfilled order, and repeat purchase using agreed windows. Compare like with like and retain costs and exceptions. The useful result of this matrix is a clearer explanation of the growth direction and uncertainty; realized growth still requires observed business records.
Check your work
- All options use the same baseline and classification rules.
- Quadrants are not probabilities or automatic priorities; uncertainty remains visible.
- Demand, capabilities and approval are explicit; no option is recorded as achieved growth.
Framework references
References explain the framework and its scope; they do not validate the fictional example's results.
Put the framework to work
Copy the outline and add your question, evidence, and open questions.