Module 6 — After the Decision
Activity 6.1
Assumption Audit
15 min
Think of a decision your team made in the last 3-6 months. It could be a programme change, resource allocation, process redesign, or strategic choice. Name the assumptions that drove it, then assess each one against what has happened since.
What did your team decide? Be specific about what was chosen and what was rejected.
Assumption 1
Assumption 2
Assumption 3 (optional)
For any assumption marked Contradicted or Still Untested: what needs to happen now?
When this decision was made, did your team define a leading indicator? Is anyone still monitoring it?
Your Audit Summary
Share Prompt
What was one assumption that turned out to be wrong, or that you realised is still untested months later?
Activity 6.2
Reflection: Momentum vs Evidence
5 min
Private reflection. This is for you, not for the room. Think honestly about an initiative that may be continuing by momentum rather than evidence.
The initiative
What would an honest assumption audit reveal?
Go / Adjust / Stop?
Activity 6.3
Trace Your Leading Indicator
20 min
Work in your Day 1 group. Pick one decision from your Day 1 submissions (Blind Spot or Decision Brief). You are going to trace backwards from the outcome to find a leading indicator you can actually act on. Use the Upstream Question Model: start at the end result and keep asking "what had to happen before this?"
Step 1: Name the Outcome
What is the lagging indicator or outcome your decision is trying to influence? This is the number that tells you what already happened. It is where most teams start and stop.
The decision (from Day 1)
The lagging indicator
What outcome are you currently watching? (e.g., Retention Rate, Active Member Share, borrowing volume, programme attendance)
Step 2: Build the Upstream Chain
Work backwards from the outcome. For each step, ask: "What had to happen before this became visible?" Keep going until you run out of upstream steps. Not every step needs to be filled. The chain is as long as your reasoning takes you.
OUTCOME
This is your lagging indicator from Step 1. You are working backwards from here.
↑
Link 1
What happened just before the outcome?
↑
Link 2
What happened before that?
↑
Link 3
What happened before that?
↑
Link 4
Further upstream (if applicable)
↑
Link 5
Root cause or earliest driver
Step 3: Circle the Earliest Measurable Signal
Look at your chain. Not every link is trackable. Find the earliest point where you have data or could get it. That is your leading indicator candidate.
Which link is your leading indicator?
Which link in the chain is the earliest point where you have data (or could realistically get it)?
What does it watch?
How often would you check it?
What action would you take if it moved?
Do you have this data today?
The Test
If this signal changed, would you have time to do something before the outcome changes? If yes, it is a leading indicator. If no, it is just an earlier lagging indicator. Go back up the chain and look for something further upstream.
Module 7 — The Full Decision Cycle
Activity 7.1
Go / Adjust / Stop: NLB Scenarios
20 min
Work in your group. Read each scenario below. For each one, discuss the evidence, identify which assumptions are holding or broken, and decide: Go, Adjust, or Stop. There is no single right answer. The discipline is in making the choice explicit and being able to explain why.
Scenario A: The Digital Personalisation Pilot
NLB launched a digital personalisation pilot on the mobile app six months ago. The goal was to increase digital visitorship by recommending content based on borrowing history.
Results so far: Digital visitorship to personalised pages is up 34%. Users who see recommendations click through at 2.8x the rate of the generic homepage. However, overall digital visitorship has only grown 3% (against a 70.5m target). The pilot is serving existing heavy users, not attracting new digital visitors. A patron privacy review flagged that the recommendation engine uses browsing data that was collected under the old consent framework.
Your group's decision
Why? What is the key evidence?
Scenario B: The Community Volunteer Expansion
To support the reopening of four revamped libraries, NLB launched a volunteer recruitment drive targeting seniors (age 50+) six months ago. The goal was to grow the active volunteer base by 15%.
Results so far: Sign-ups exceeded target by 40%. However, only 22% of new sign-ups have actually volunteered for any activity. The inactive rate among senior volunteers is significantly higher than for younger cohorts. Exit surveys (small sample) suggest scheduling conflicts and unclear role expectations. Meanwhile, the four libraries reopen in eight weeks and staffing gaps remain.
Your group's decision
Why? What is the key evidence?
Scenario C: The Quarterly Report Consolidation
Nine months ago, the Data Office consolidated twelve divisional quarterly reports into a single cross-NLB dashboard. The goal was to reduce reporting burden and improve strategic visibility across the 10 topline KPIs.
Results so far: The dashboard was built on time and on budget. It is technically functional and refreshes monthly. However, analytics logs show only 14 unique users access it per month (out of approximately 80 data leaders and stewards). Six divisions have quietly resumed producing their own reports alongside the dashboard. When asked, division heads say they need "more granular data" than the dashboard provides. The Data Office has spent 30% of one FTE maintaining the dashboard.
Your group's decision
Why? What is the key evidence?
Reflection
Which scenario was hardest to decide? That difficulty usually means the evidence is genuinely mixed and reasonable people could disagree. That is when the discipline of making the choice explicit matters most.
Activity 7.2
From Decision Brief to Tool-Ready Question
25 min
Work in your Day 1 group. Pull up the Blind Spot Documentation, Decision Brief, and Data Requirements you created on Day 1. You will use them to prepare a structured question for the Qlik tools this afternoon. Each step builds on the previous one.
Round 1: Classify Your Question
Look at your Decision Brief. Which inquiry mode does your group's question fall into?
Your Decision Brief question (from Day 1)
Inquiry mode
Round 2: Sharpen the Prompt
Rewrite your Decision Brief question as a specific prompt for each tool. Use the starters below to help structure your question.
Qlik Answers Prompt
Ask a descriptive or diagnostic question in plain language. Be specific enough that the tool can return a direct answer.
Starter patterns:
"What is the [metric] for [group A] vs [group B], broken down by [dimension]?"
"Which [category] has the highest/lowest [outcome] over the past [time period]?"
"How has [metric] changed since [event or date], segmented by [dimension]?"
"Of [specific patron segment], what percentage [took a specific action] within [time window]?"
"What is the [metric] for [group A] vs [group B], broken down by [dimension]?"
"Which [category] has the highest/lowest [outcome] over the past [time period]?"
"How has [metric] changed since [event or date], segmented by [dimension]?"
"Of [specific patron segment], what percentage [took a specific action] within [time window]?"
Qlik Predict Scenario
Frame a predictive or prescriptive question using one of your Decision Brief options. What would you like to forecast?
Starter patterns:
"If we [take action from Option A], what is the projected change in [outcome] over [time period]?"
"What are the key factors driving [outcome]? Which ones can we influence?"
"If [leading indicator] drops below [threshold], what is the predicted impact on [KPI]?"
"Compare the projected [outcome] for Option A vs Option B vs Option C over [time period]."
"If we [take action from Option A], what is the projected change in [outcome] over [time period]?"
"What are the key factors driving [outcome]? Which ones can we influence?"
"If [leading indicator] drops below [threshold], what is the predicted impact on [KPI]?"
"Compare the projected [outcome] for Option A vs Option B vs Option C over [time period]."
Round 3: Stress-Test with Your Blind Spot
Before you use any tool, know what it can and cannot tell you.
Must-have data connection
If the connection does not exist...
Limitation statement
Write the exact sentence you would include in a recommendation to caveat what the tool cannot see.
Quality Check
Look at your Qlik Answers prompt. Could someone outside your team understand exactly what you are asking? If not, make it more specific. Look at your Predict scenario. Does it name a concrete action, a measurable outcome, and a time horizon? If not, sharpen it.