BehavioralStrategy.com teaching pack v1.0
Author: Jason Hreha
Source file: education/verify-decision.md
License for original text: CC BY 4.0. External sources retain their rights.
Editable text copy; Markdown headings and tables are retained.

# Verify: Define Success

> **Your decision:** What should a service team investigate next, and which completed action should determine whether that work helps users?

This exercise starts **after the National Audit Office published its Verify investigation on 5 March 2019**. The audit findings are evidence available at the decision point. Do not use them as information that an earlier team necessarily had. Allow about 20 minutes as a teaching estimate.

Use the [decision workbook](https://behavioralstrategy.com/education/decision-workbook/) or a blank page. The task is a public-source reconstruction for learning [Behavior Market Fit testing](https://behavioralstrategy.com/methodology/validate-behavior-market-fit/), not a description of the government's actual next decision.

## The public record

Verify used commercial providers to confirm identities for access to government services. The NAO's March 2019 investigation reported the following. [Evidence BS-0081; see source-notes.txt]

| Reported item | Boundary stated by the source |
|---|---|
| Verification success was **48% in February 2019**. | The measure concerned successful Verify sign-up in a single attempt under the source’s counting rules. |
| The earlier projection was **90%**. | This was a planning expectation, not an experimental control group. |
| Applicants could leave before choosing an identity provider. | Those departures were excluded from the success measure. |
| Verification could succeed without access to the intended service. | The reported rate did not establish completion of the user's government task. |

The audit also reported additional manual work for departments. It did not isolate one cause of the program's difficulties. [NAO findings](https://www.nao.org.uk/press-releases/investigation-into-verify/) and [full report, paragraphs 1.16 and 1.17](https://www.nao.org.uk/wp-content/uploads/2019/03/Investigation-into-verify.pdf#page=18).

## Your role and boundaries

**Instructor-created scenario:** you advise the owner of one online government application. The team wants an eligible person to submit the application and receive confirmation that it has been accepted for processing. Identity assurance remains a required condition. An eventual approval, payment or other service benefit is a later outcome.

You may recommend one bounded investigation or pilot for that existing service. Government-wide funding and the identity-assurance standard are outside your decision. You have no additional service-specific funnel, cost or user-research data. If a number is missing, mark it as missing.

## Repair the claim before choosing an intervention

This is a **fictional briefing statement written for the exercise**:

> “Verification success is 48%, so about half of people complete the government task. Increasing verification success will show that we are meeting their need.”

Use [workbook step 1](https://behavioralstrategy.com/education/decision-workbook/#step-1) to separate the user's useful outcome, the actions required and the intermediate system event. Write a replacement statement that preserves what is supported and names what is unknown.

Then use [workbook step 6](https://behavioralstrategy.com/education/decision-workbook/#step-6) to explain why subtracting the reported rate from the projection does not estimate the effect of an intervention. A performance shortfall and a causal effect are different claims.

## Map the whole journey

Start before the person selects an identity provider and continue beyond verification to the application confirmation. Use [workbook step 3](https://behavioralstrategy.com/education/decision-workbook/#step-3).

For each step, record:

- Who acts: applicant, identity provider, service team or support staff.
- What observable event shows progress.
- What conditions or help are necessary.
- Whether an unsuccessful step means refusal, inability, a technical failure, a competing route or an unknown result.

Include repeat attempts and existing accounts. A person who tries again should not silently become a new person in the denominator. A person who uses another permitted route may still achieve the intended outcome.

## Choose the next priority

These are original teaching options, not a reconstructed government shortlist.

| Option | Immediate aim | Evidence you would need |
|---|---|---|
| Increase awareness | Bring more intended users into the route. | Whether additional entrants can complete a useful task under current conditions. |
| Improve registration | Help people complete the identity-provider step. | Where registration fails and whether success carries through to the service. |
| Repair the service handoff | Help an already verified person reach and complete the application. | Whether account, record-matching or other transition problems block the task. |
| Reduce or support the required work | Reuse permitted existing steps or provide a sustainable assisted route. | What work disappears or moves to staff, the resulting completion and the cost of support. |

You may first recommend an investigation that distinguishes these options. Use [workbook step 4](https://behavioralstrategy.com/education/decision-workbook/#step-4) to explain why your choice is more useful than simply pursuing a higher intermediate percentage.

## Specify a measurement plan and decision rule

Complete the following fields before reading the worked response.

| Field | What your record must specify |
|---|---|
| Population and episode | Which applicants and which application journey are included? |
| Start and finish | What event begins the observation, and what confirms the useful task is complete? |
| Denominator and window | Are you counting people, attempts or applications, and for how long? |
| Observation gaps | How are early departures, missing links, repeat attempts and unresolved outcomes recorded? |
| Support and alternatives | Which help and other routes remain available, and what do they require from staff? |
| Next comparison | What would you need to compare before claiming that a proposed change improved completion? |
| Decision rule | Which finding leads to a repair, a different route, a larger test or an inconclusive result? |

Use [workbook step 7](https://behavioralstrategy.com/education/decision-workbook/#step-7) to check the diagnosis. An unsuccessful treatment does not by itself prove that applicants cannot perform the target behavior. Equally, a successful identity check does not establish the whole journey's feasibility.

## Submit your recommendation

Your record should contain a repaired briefing statement, a complete action chain and a recommendation for the next commitment. Name the strongest alternative and the observation that would make you switch to it. State which claim you would decline to make from the current evidence.

## Source chronology

| Observation period | Public record | Evidence role |
|---|---|---|
| Program performance through February 2019 | [NAO investigation, 5 March 2019](https://www.nao.org.uk/reports/investigation-into-verify/) | Independent government audit, available before this exercise's decision. |
| The same investigation | [NAO findings release, 5 March 2019](https://www.nao.org.uk/press-releases/investigation-into-verify/) | Summary of the same audit, with no separate replication. |

Sources checked **5 September 2026**. No later program outcome is needed to answer this exercise. The source record does not show use of DRIVE, BMF or BFA.

After writing your record, open the [Verify worked decision](https://behavioralstrategy.com/education/verify-worked-decision/). Use the [measurement lab](https://behavioralstrategy.com/education/measurement-lab/) if your counts or observation rules remain unclear.

*Original teaching material by Jason Hreha, version 1.0, licensed CC BY 4.0. External sources retain their own terms. See [teaching notes and reuse guidance](https://behavioralstrategy.com/education/teaching-notes/).*
