How we decide
Carrier-route fit scoring computed only from NCRD route records served by MailEastCoast. USPS deliverable counts are measured; every demographic input is modeled from ACS block groups via a block-areal-weighted crosswalk and carries its own confidence range. Estimate precision grades the score: the wider a contributing range, the lower the route scores.
What we measure
| Component | Weight | What it means |
|---|---|---|
| Deliverability | 0% | 0.50 x residential share of deliverable addresses (res_cnt / tot_cnt) + 0.30 x pct_owner_occ + 0.20 x (100 - vacancy share, from est_hu_vacant / est_hu). Deliverable counts are measured USPS figures. |
| Audience fit | 0% | Mean of trapezoid fit scores for median_hh_income and median_home_value against the target bands of the selected profile. 100 inside the ideal band, sloping to 0 at the outer bounds. |
| Household fit | 0% | Mean of trapezoid fit scores for pct_with_kids, pct_65_plus and avg_hh_size against the profile target bands. |
| Reach quality | 0% | 0.35 x pct_vehicles_2plus + 0.25 x pct_health_ins + 0.20 x (100 - pct_lang_non_eng) + 0.20 x (100 - pct_wfh scaled), all modeled ACS estimates. |
Where the line sits
- 75+ — Prime
- 60+ — Strong
- 45+ — Marginal
- below 45 — Cut
Reasons we cut a route
- LOW_DELIVERABILITY_MIXPoor delivery mix
- AUDIENCE_MISMATCHWrong audience
- BELOW_COST_THRESHOLDCosts more than it returns
- LOW_CONFIDENCEData too imprecise to justify
- INSUFFICIENT_DATAProvenance incomplete
Retired reasons, kept readable on older plans: [object Object], [object Object].
Scoring profiles
Balanceddefault
No vertical assumptions. Use this when you just want the obviously wasted routes cut.
deliverability 0% · audience_fit 0% · household_fit 0% · reach_quality 0%
Dental and local health practice
Insured households within a short drive, family-weighted, mid to high income.
deliverability 0% · audience_fit 0% · household_fit 0% · reach_quality 0%
Furniture and home goods
Higher home value and discretionary income; new and move-up homeowners.
deliverability 0% · audience_fit 0% · household_fit 0% · reach_quality 0%
Home services and roofing
Owner-occupied single-family households that can authorise work on the building.
deliverability 0% · audience_fit 0% · household_fit 0% · reach_quality 0%
Restaurant and local retail
Dense, everyday-spend households close to the storefront.
deliverability 0% · audience_fit 0% · household_fit 0% · reach_quality 0%
Senior services
Older households, home-based, less reliant on digital channels.
deliverability 0% · audience_fit 0% · household_fit 0% · reach_quality 0%
What the data is
Address counts per carrier route are measured USPS delivery statistics. Every demographic input is modeled onto the route from census block groups and is published with the confidence range around it. Where that range is wide, the route's score is pulled toward neutral rather than trusted at face value. Nothing here is a measured response rate.
Covered ZIPs today: 28202, 28203, 28204, 28205, 28206, 28207, 28208, 28209, 28210, 28211, 28212, 28213, 28214, 28215, 28216, 28217, 28226, 28227, 28262, 28269, 28270, 28273, 28277, 28278
Where coverage comes from
We do not decide which counties are complete. Each county below carries the route data feed's own completeness counters, read at the time shown. A county that stops reporting zero on both counters drops out of coverage on its own, and any ZIP inside it is withheld rather than scored.
- Mecklenburg County, NCcomplete
22 routes served · 0 missing covariate basis · 0 missing ordered range · checked 9/13/2026, 10:02:55 PM · /37/119?zip=28270
Version history
- v2.0-ncrdmodeled9/4/2026Carrier-route fit scoring computed only from NCRD route records served by MailEastCoast. USPS deliverable counts are measured; every demographic input is modeled from ACS block groups via a block-areal-weighted crosswalk and carries its own confidence range. Estimate precision grades the score: the wider a contributing range, the lower the route scores.
- v1.0-modeledmodeled9/4/2026Baseline Don't-Mail-Here score. Four components computed from modeled route-level demographic and behavioural inputs, combined with published weights. Nothing in this version is measured response data; every score is labelled modeled.