intermediate
2 min answer
A fast-growing company has dozens of internal applications of unknown value. How should the portfolio be assessed, and what decisions should follow?
Show the full answer Hide the answer
The two axes that matter
Business value and technical condition, assessed independently. Four quadrants follow, and each implies a different decision:
| High business value | Low business value | |
|---|---|---|
| Good technical condition | Invest — extend and build on it | Tolerate — leave it alone, spend nothing |
| Poor technical condition | Migrate or rebuild — this is where investment belongs | Retire |
The retire quadrant is the one organisations avoid, because someone still uses it and switching them off requires a conversation. It is also the highest-return quadrant, since it removes cost, risk and cognitive load permanently.
How to assess honestly
- Business value from usage and outcomes, not from stakeholder opinion. Who uses it, how often, and what breaks if it stops.
- Technical condition from evidence: change failure rate, incident rate, time to make a change, dependency currency, whether anyone still understands it.
- Run cost including the people cost, which is usually the larger half and is frequently uncounted.
- Include shadow systems. A spreadsheet that a business process depends on is part of the portfolio, and it is usually the highest-risk item because it has no owner and no backup.
The decisions that follow
- Retire ruthlessly. Every retirement removes permanent cost and risk.
- Invest only in the high-value quadrant, and be explicit that low-value systems in poor condition receive nothing regardless of who complains.
- Tolerate deliberately. A system that is unpleasant and works and is cheap to run does not need fixing, and saying so protects capacity for what does.
The failure to avoid
Assessing and doing nothing. A portfolio assessment that produces a document rather than a set of retirement decisions has cost effort and changed nothing.
The output should be a small number of specific decisions with owners and dates, and the assessment should be scoped to what can actually be acted on in a quarter rather than to comprehensive coverage.