Tracking Roadmap
Nothing on this page is a finding. It lists what the current exports cannot answer and the exact data needed to unlock it — regenerated from each upload, so a repaired tracking defect disappears here automatically.
Configurator selections
8,219
All filter selections in the current configurator upload.
Observedn = 8,219
Usable for analysis
94%
534 selections carry no analysable value and are discarded.
Observedn = 8,219
Blocking data gaps
3
Gaps that make a whole analysis family impossible today, not just less precise.
Observedn = 9
Leads with unusable budget
62
Of 2,496 leads — form validation would remove this loss.
Observedn = 2,496
What the configurator export delivers today
Per filter type, counted from the current upload.
3 rows
| Filter type | Selections | Distinct values | Used | Reason |
|---|---|---|---|---|
| Size | 4,683 | 60 | Yes | 60 distinct option values, mapped and used in the analysis |
| Occupancy | 3,002 | 2 | Yes | 2 distinct option values, mapped and used in the analysis |
| Term | 534 | 1 | No | Every selection exports without a value — tracking defect |
Priority 1 — blocks whole analysis families
2 items.
2 rows
| Analysis unlocked | Required data / export | Grain | Status in the current upload |
|---|---|---|---|
| Configurations, features, FF&E ranking, feature co-occurrence, true demand clusters | Configurator event export with one row per selected option, including feature groups (bed, desk, storage, kitchen, bathroom, seating, add-ons) | one row per session × option | Today the export delivers 3 filter types (Size, Occupancy, Term) and no feature-level data. |
| Price × size × product × timing on user level, and the search → configurator → lead funnel | A shared session / visitor id in all three exports | one id per visitor session | No shared key exists today: searches and configurator rows carry no id, 2,496 of 2,496 leads carry a hashed contact key only. All cross-source links stay at location level (Indication). |
Priority 2 — materially improves precision
4 items.
4 rows
| Analysis unlocked | Required data / export | Grain | Status in the current upload |
|---|---|---|---|
| Configurator demand over time, seasonality of product preference, period filtering of product KPIs | Timestamp on every configurator selection | one timestamp per selection | The configurator export has no dates, so product pages ignore the period filter. |
| Term as an analysable dimension (534 selections currently discarded) | Repair the Term tracking so the selected value is written to the export | one value per selection | Every selection exports without a value — tracking defect |
| Realised willingness to pay, conversion per price band, ERV validation | Booking / contract export: property, room type, size, contract rent, contract start, term, booking date | one row per contract | No booking data in the app — every price figure is stated willingness to pay, never a realised rent. |
| Clean budgets without manual plausibility filtering | Form validation on the exit-intent budget field (200–3000 EUR per month) | — | 62 of 2,496 leads carry an implausible or empty budget and are excluded. |
Priority 3 — adds new dimensions
3 items.
3 rows
| Analysis unlocked | Required data / export | Grain | Status in the current upload |
|---|---|---|---|
| Search-term intelligence: amenities, university, transport, furnished, private bathroom, shared living | Site-search keyword export: term, timestamp, location, result count | one row per query | Only date-range searches are exported today; free-text queries are not captured. |
| Origin × price × room type × stay | Country of origin or university field in the exit-intent form (or aggregate IP country) | one attribute per lead | No origin attribute in the lead export. |
| Room-type demand (studio / en-suite / shared) and hard unit-mix evidence | Room-type attribute on configurator sessions and leads | one attribute per session / lead | Occupancy (single vs double) is the only product signal available today. |
How to commission this
Each row above is written so it can be handed to the website or CRM team as a specification.
- The single highest-value change is one shared session id across searches, configurator and exit-intent form — it turns every location-level indication into a user-level fact.
- The second is a booking export, which converts stated willingness to pay into realised rent and validates the ERV assumptions (Estimated Rental Value — the rent a room is expected to achieve).
- Everything else adds dimensions (features, room type, origin, keywords) rather than changing the reliability of what is already here.