Before you walk a street, know why you picked it.
Explore neighborhoods by income, homeownership, housing density, and audience fit. Combine that view with your existing coverage to decide where your next campaign should go.
- Census lenses for income, density and home ownership
- A fit score for each area
- The coverage you already have, on the same map
Private beta. A fit score is a planning signal, not a sales forecast.
Every score is arithmetic you can check.
Recommendations rank a city’s neighborhoods by four census signals, blend in your audience when you have described one, then discount the ground you have already walked. Move the inputs and follow the working. The weights and thresholds are the ones in the app. The input values are sample data.
Inputs
Strength is each area’s position between the lowest and highest neighborhood in the city, from 0 to 100.
Working
| Signal | Strength | Base weight | Weight used | Adds |
|---|---|---|---|---|
| Median income | 0.78 | 0.35 | 0.35 | 0.27 |
| Owner-occupied share | 0.84 | 0.30 | 0.30 | 0.25 |
| Density | 0.31 | 0.20 | 0.20 | 0.06 |
| Door count | 0.62 | 0.15 | 0.15 | 0.09 |
| Opportunity | 0.68 | |||
- Blend in audience fit
0.68 × 0.65 + 0.81 × 0.35 = 0.73 - Discount covered ground
0.73 × (1 − 0.04) = 0.70 - Round to a score out of 100
70
- high income
- matches your audience
- untouched
Up to three reasons appear beside each recommendation, and the coverage reason is always one of them. The app lists the five highest scores for your city.
Five lenses on the same map.
The best neighborhood for a mobile detailer may not be the best one for a new pizza shop. Each lens shades the city by one question, so you can switch the question instead of squinting at one blended colour.
| Lens | What it shades by | Map legend |
|---|---|---|
| OpportunityThe default view | Income, owner-occupied share and density, blended 40 / 35 / 25 | Low, mid, high |
| Income | Estimated median household income | $, $$, $$$ |
| Doors | Dwelling count | Low, mid, high |
| Homeowners | Owner-occupied share of dwellings | Low, mid, high |
| FitNeeds an audience first | Match to the customer you described, 0 to 100 | Low, mid, high |
Shading splits into thirds at 0.33 and 0.66 of the range between the city’s lowest and highest neighborhood. An area with no value for the active lens is left unshaded.
Describe the customer, and Fit starts scoring.
Pick any of four traits. Leave a trait blank and it is ignored. Each chosen trait scores an area from 0 to 1 by where it ranks in the city, and fit is the average of the traits you chose, shown out of 100.
| Trait | Choices | How an area earns it |
|---|---|---|
| Tenure | OwnersRenters | Rank of the owner-occupied share. Renters flips the rank. |
| Income | LowMidHigh | Rank of median income. Mid rewards areas closest to the middle of the city. |
| Density | SuburbanUrban | Rank of people per square kilometre. Suburban flips the rank. |
| Household | SmallLarge | Rank of people per dwelling. Small flips the rank. |
- No traits chosen
- Fit sits at a neutral 50
- Areas with no dwellings
- Excluded, there is nobody to flyer
- Pull on a recommendation
- 35% fit, 65% opportunity
Use the ranking as a planning aid, together with what you already know about your customers and your service area.
The reasons printed beside a recommendation
Every recommendation says why it is on the list. These are the only reasons the app can give, and the exact line each one has to clear.
- high income
- Income strength of 0.66 or more
- solid income
- Income strength from 0.40 up to 0.66
- owner-occupied
- Six in ten dwellings or more are owned by the people living in them
- lots of doors
- Door-count strength of 0.66 or more
- dense
- Density strength of 0.66 or more, shown when door count did not already qualify
- matches your audience
- Audience fit of 66 or more
- untouched
- 5% of the area or less overlaps your coverage
- barely covered
- More than 5% and under 40% covered
- partly covered
- From 40% covered up to the 80% cut-off
Where the numbers come from.
The private beta is gated to Edmonton, with 287 neighborhoods loaded from public open data. Each region records its source and census year, so the app can tell you how recent a figure is.
- 287 Edmonton neighborhoods
- 6 fields per area
- 1 city in the beta
| Field | Source | Year | Read it as |
|---|---|---|---|
| Population | City of Edmonton Open Data | 2021 federal census | Used for density and for people per dwelling. |
| Median household income | City of Edmonton Open Data, income bands | 2016 census | An estimate. The city publishes income bands, so the median is interpolated from the band counts. |
| Dwellings | City of Edmonton Open Data, tenure counts | 2012 | Owned plus rented dwellings. This is what the app calls doors, and it is a count of homes on record. |
| Owner-occupied share | Same tenure counts | 2012 | Owned dwellings divided by all dwellings. |
| Density | Population over the published land area | 2021 population | People per square kilometre. |
| Location | Published neighborhood centre points | Current open data | An approximate centre. An area with no centre point cannot be measured for coverage and reads as untouched. |
To measure coverage against an area, your covered ground is stored as circles, and the app sums how much of them overlaps the neighborhood’s footprint, capped at 100%. The footprint is itself an estimate, a circle sized from population and density and held between 200 m and 3 km in radius.
A good area still needs a workable plan.
- 1
Move from an area to a campaign.
Keep the target neighborhood and the campaign together so the team knows what the drop is meant to reach.
- 2
Divide the work for your crew.
Route cutting partitions a campaign area into territories. Balancing can use area or available dwelling estimates, and the planned guide gives the walk a starting point.
See crew planning - 3
Use outcomes to inform the next decision.
Review scan and conversion activity alongside coverage. The longer-term aim is to learn which kinds of areas produce revenue for your business. Today’s opportunity scores are explainable heuristics, not a trained sales-prediction model.
See campaign results
A little more detail.
- Does the model predict my sales?
- Current recommendations use weighted neighborhood characteristics, audience fit, and coverage. Learned revenue-per-flyer prediction is part of the longer-term direction, not a promise made by today’s scores.
- Can a score tell me exactly which household will buy?
- No. These are area-level planning signals. They do not identify an individual household’s likelihood of becoming a customer.
- Why did two areas with similar incomes get different scores?
- Income is one of four signals, and coverage is applied last. An area you have half covered keeps only half of its blended score, so untouched ground with the same census profile ranks above it.
- What happens when the census has no figure for an area?
- The missing signal is dropped and the remaining weights are scaled up to fill the gap. Missing fields stay missing on the card. They are never filled with a guess.
Know where your next 500 flyers should go.
Scout your own neighborhoods before you print anything.
- Join the list
- Get your invite by email
- Walk your first street