Mapping the public sector funnel from awareness to sustained action

Movember asks for one action a year. Transport for London asks for one twice a day and gets it without asking. Same funnelfunnelThe customer journey from awareness to purchase, typically Awareness, Interest, Consideration, Decision, Action, with prospects narrowing at each stage.View full definition → shape, opposite failure points. A stage mapmapUsing software to automate repetitive marketing tasks and campaigns, enabling personalisation at scale across channels like email, web, and social.View full definition → exists to tell you which gap you are actually looking at, because a single top-to-bottom conversion rateconversion rateThe percentage of visitors or prospects who complete a desired action (purchase, sign-up, contact form), calculated as conversions divided by total opportunities.View full definition → cannot distinguish a campaign nobody saw from a campaign everybody saw and abandoned at the form.

Six stages, from unaware to habitual

Name them so the boundary between any two stages is arguable in a meeting. Fuzzy labels produce dashboards nobody trusts.

1. Unaware. The share of the eligible population that has never met the message. You never measure this directly; you infer it as eligible population minus measured awareness, which is why the denominator argument below matters more than it looks.

2. Aware. Can recall the message, prompted or unprompted. ImpressionsImpressionsThe total number of times an ad or piece of content is displayed, regardless of clicks. Each display counts as one impression, even to the same person.View full definition → are not awareness; they count deliveries. Movember's awareness stage is odd, because the medium is a moustache on a colleague's face: awareness is manufactured by people who already converted.

3. Receptive. Accepts that the message applies to them and that the messenger is credible. Distrust and "not for people like me" both stall here, and neither responds to more reachreachThe number of unique people exposed to your message in a given period. Unlike impressions, reach counts each person once, no matter how often they see it.View full definition →.

4. Intent. A low-commitment step with a timestamp: starts the form, registers a fundraising page, requests the badge, books the slot.

5. First action. The behaviour itself. Files the return, gets the jab, taps in with a contactless card, sets the office to 28C and drops the tie, sends the first donation.

6. Habitual. Repeats without being prompted again.

That last boundary is where public sector reporting quietly cheats. Someone who gives a second time because a second mailer arrived has repeated an action; they have not formed a habit. Keep them in stage 5 and count the prompt as a cost. Habit means the behaviour survives the withdrawal of your marketing, which is the only version that scales without a permanent budget line. TfL, which accepted contactless bank cards on buses from December 2012 and across the Tube from September 2014, reaches stage 6 within a handful of journeys, because the behaviour repeats daily and the alternative is a queue. Movember rebuilds stages 2 to 5 every October.

The denominator argument

Before any rate means anything, decide who is in the funnel at stage 1.

• TfL can count journeys or travellers. Journeys inflate the base and flatter the top of the funnel; travellers are harder to count and are the only unit that makes a retention rate honest.
• Movember's eligible base is not "men". It is men in a country where the campaign runs who sit inside a plausible recruiting network (workplace, sport, student), plus the Mo Sisters who fundraise without the moustache.
• Cool Biz, run by Japan's Ministry of the Environment since June 2005, has a base of workplaces, not workers. The ministry asked offices to hold air conditioning at 28C and let staff work without jackets and ties through the summer. An individual could not comply until the employer allowed it.

Pick the wrong denominator and you can halve or double every rate in the deck without a single real behaviour changing. Put the definition on the first slide.

Where the drops happen and what each one means

• Unaware to aware. Usually the largest percentage loss in the funnel. Paid reach is passive and nobody is shopping for your service. Cool Biz got round this in 2005 by having cabinet ministers appear on television open-necked, which converted a policy request into free editorial coverage.
• Aware to receptive. A trust and relevance gap, not an information gap. The tell: recall is high, engagement is flat. More spend behind the same message makes it worse.
• Receptive to intent. Complexity. Eligibility rules people cannot self-assess, document lists, forms demanding information the applicant has to go and find. US benefits enrolment portals have been flagged repeatedly by the Government Accountability Office for abandonment tied to exactly this.
• Intent to action. Friction with a deadline attached: appointment availability, expired ID, transport, opening hours. The cheapest wins in the sector sit here. The UK's Behavioural Insights Team has published trial after trial where a message sent at the right moment moved completion by double-digit percentage points at a cost per person close to nothing.
• Action to habitual. Chronically weak, and the sector's own figures say so. Fundraising Effectiveness Project reporting has put overall donor retention in the mid-40s and first-time donor retention around 20 percent for years (directional, and the methodology shifts). Four in five first-time donors never come back, so a funnel that treats the first gift as the finish line rebuilds its base annually.

Nested funnels multiply, they do not add

When the converting unit is an organisation and the behaviour is performed by individuals inside it, you have two funnels stacked and the rates multiply. Cool Biz has an adoption funnel (does this workplace move the thermostat and relax the dress code?) and a compliance funnel (do the people in it actually work tieless at 28C?). 90 percent adoption with 40 percent individual follow-through gives 36 percent of the population outcome, not 90.

Two consequences. Reporting the organisational number alone is the commonest way public campaigns overstate results, and it is rarely deliberate: the organisational number is simply the one that is easy to collect. And at stage 6 you often cannot observe the behaviour at all. Cool Biz uptake has been tracked largely through employer and worker surveys, so the habitual stage is self-reported. If the only evidence for your best stage is a survey, put that in the footnote.

The one piece of arithmetic the map needs

Stage conversion rate = (Number who complete stage) / (Number who entered stage) × 100

Compute it per boundary, never only end to end. An overall 0.5 percent unaware-to-action rate can hide a healthy 60 percent intent-to-action rate sitting behind a 2 percent awareness rate. The two problems share nothing: one is a media buy, the other is a form.

The map is also where the module's other numbers attach. What each conversion costs belongs to the cost-per-action lesson; what a habitual actor is worth belongs to the lesson on valuing people who never pay you directly; who you are allowed to compare yourself with belongs to the peer-set lesson. Here, compare a stage against the same stage of your own previous campaign, which is the only comparison the map by itself justifies.

Two failure modes in reading the map

Improving the funnel by shrinking it. Target only the already-receptive and every rate below stage 2 climbs. Reach falls, conversion rates look excellent, population outcome gets worse. If stage rates rise while absolute actions fall, you have optimised against the mission. Show the absolute count beside every percentage.

Treating drop-off as zero. In a commercial funnel a lost prospect costs you the acquisition spend. In a compliance funnel the loss is negative: a non-filer becomes an enforcement case, a missed vaccination slot goes unfilled, an abandoned application returns as a call to a contact centre. Drop-off at stage 5 carries a price tag, usually larger than the marketing that would have prevented it.

A third trap belongs to seasonal campaigns. Annual behaviour cannot be measured on a 30-day lapse window. Movember's retention question is answered eleven months after the action, and a dashboard built on monthly cohorts will report a catastrophe every December.

🎬 [VIDEO: "How the UK Nudged a Nation" - youtube.com - a look at Behavioural Insights Team case studies showing how small design changes moved people from intent to completed action in government services]

Instrumenting the stage boundaries

Basic event tagging is enough. One row per boundary crossing, with the stage named:

event_name, user_id, timestamp, stage
"ad_impression", u1023, 2026-02-01T10:03, reach
"site_visit", u1023, 2026-02-01T10:05, receptive
"form_started", u1023, 2026-02-01T10:07, intent
"form_submitted", u1023, 2026-02-01T10:12, action
"renewal_unprompted", u1023, 2027-02-03T09:00, habitual

Two rules make this usable: one event per boundary rather than per click, and a written stage definition that survives a change of agency. Notice what is missing. Awareness comes from survey work, not the event stream, and the unaware row cannot exist at all, which is why the population estimate lives outside this table and gets cited every time a rate is quoted.

Key Takeaways

• Six stages: unaware, aware, receptive, intent, first action, habitual. The last boundary is the one teams fudge, and a prompted repeat is not a habit.
• Every rate depends on the denominator you declare at stage 1 (journeys or travellers, workplaces or workers). Fix it in writing before computing anything.
• Each drop diagnoses something specific: reach, then trust, then complexity, then friction, then stewardship. First-time donor retention around 20 percent (Fundraising Effectiveness Project, directional) marks where the sector's weakest boundary sits.
• Nested funnels multiply. 90 percent organisational adoption times 40 percent individual follow-through is 36 percent, and quoting the 90 is how good campaigns end up overstated.
• Rising stage rates with falling absolute actions means you shrank the funnel. And in compliance funnels, drop-off costs more than nothing: the person reappears as an enforcement case or a contact-centre call.