Problem Diagnosis and Intervention Logic · Lesson 3

Compare feasible intervention alternatives

6 min read

Lesson objectives

  • Apply compare feasible intervention alternatives to a development case, explaining inputs and decision limits.
Reading tools

Once the problem is clear, teams often jump to the intervention they know how to deliver. If the bottleneck is employer interviews, a longer training course is not necessarily the answer. Construct alternatives with different mechanisms: match jobseekers to verified vacancies, offer a short employer-designed practical course, provide limited travel support for interviews, or continue the existing service while improving measurement. The last option is a decision baseline, not always ‘doing nothing’; it has continuing costs and outcomes. Describe each option's dose, population, period, resources, and delivery conditions so that a vague title is not compared with a fully specified program. Do not place an option that is legally or operationally unavailable into a weighted ranking before resolving its constraint. A fundamental eligibility condition is a gate, not just a low score that a high projected benefit can offset.

Begin with a conditions matrix rather than a score matrix. Ask whether beneficiary data can be used for this purpose, whether a partner can deliver during the period of need, and whether outcomes and complaints can be followed without exposing sensitive information. If an answer is unknown, record a verification test and its cost instead of ticking ‘yes.’ Then compare eligible alternatives using defined criteria: potential reach, ability to remove the diagnosed constraint, resource cost, speed of learning, reversibility, and distribution of benefits and burdens. Do not count the same fact twice under labels such as ‘ease of implementation’ and ‘readiness’ when both reflect the same partner. A weighted sum does not make judgment objective merely because the weights total one. Record who chose the weights and why, and test whether a plausible weight change reverses the result. If a narrow ranking flips, the right output may be a field test of the decisive uncertainty rather than full funding for the arithmetic winner.

Alternatives are not always independent. A matching option may need brief training for a subset of jobseekers; travel support may achieve little if no interviews exist. Draw a dependency map showing what must happen before each option and what can be combined. Avoid counting the same benefit twice: if matching raises interview access, then the same additional interviews cannot be booked again as a separate travel-support benefit. Distinguish intermediate from final outcomes and calculate economic benefit incrementally against the specified alternative. If two options reach different groups, cost per beneficiary alone need not settle the decision. A funder may choose a costlier option with fairer reach and defend that trade-off explicitly. In every case, show who bears participation and travel time even when that burden is absent from the funder's budget.

Check the credibility of an interview count before entering it in a comparison. ‘We expect 90 interviews’ differs from 90 written appointments with named employers and dates, and both differ from 90 interviews actually completed. Classify evidence as intention, verifiable commitment, or completed event rather than summing the three in one column. Check whether one person may have several interviews; interviews and people are not always interchangeable units. If travel costs appear to block participation, ask whether travel is truly the constraint or whether suitable local vacancies are absent. A small set of interviews with candidates and employers may show that changing interview times is cheaper than paying transport. The alternative should respond to the diagnosed constraint rather than to an available budget line.

In the decision brief, present the comparison as an auditable argument: objective, reference alternative, constraints, evidence behind each estimate, and what finding would pause or change the choice. Express alternative costs over one horizon and in one currency, separating the funder’s expense from beneficiary and partner time; detailed economic appraisal comes later. A six-month trial may look cheaper than a year-long service only because its horizon is shorter. Nor should the largest reach count automatically win when service quality or access for those in greatest need differs. Two measures can remain side by side instead of being squeezed into one score: cost per eligible person reached and likelihood of a documented interview, with group differences noted. That makes the trade-off visible without pretending that an alternative has already produced impact.

Lesson summary

Compare feasible intervention alternatives