4. Expanded Feature Catalog (What Agencies and Contractors Ask For)
These are features home-services agencies, their contractor clients, and agencies in general ask for in reporting tools. They're each classified with the same honesty as the stated idea. Being in demand doesn't mean you should build it. Most of these demands are already met by someone.
Demand signals here are analyst judgment from the competitor feature lists read on 30 Sept 2026 (research/10), not from interviews. The first 15 calls in plans/08 should confirm or kill each P0 item.
4.1 Data sources (connectors), in priority order
| # | Source | Why it matters for home services | API reality | Priority |
|---|---|---|---|---|
| 1 | Google Ads, including Local Services Ads | Search and LSA are the main paid channels for contractors (assumed, unverified share) | Google Ads API. local_services_lead gives lead type, service category, lead status, charge status, credit status and timestamps (plus local_services_lead_conversation). Developer token: Explorer 2,880 ops/day on production accounts, Basic 15,000/day (brand verification), Standard unlimited (manual audit, ~10 business days). | P0 |
| 2 | CallRail | Most calls in this niche are tracked here (assumed) | Public API (not checked this run). Read calls, source, duration, answered or missed. | P0 |
| 3 | Jobber | The booked-job and revenue source for small and mid contractors | GraphQL, OAuth 2.0, date-versioned. Clients, requests, quotes, jobs, invoices, visits. A query-cost budget plus a 2,500-requests-per-5-minutes guard (secondary source). App Marketplace for listing. | P0 |
| 4 | Housecall Pro | The other main field-service system for small contractors | REST API with jobs, customers, estimates, invoices, payments, webhooks, and lead_source on jobs. Full API access is tied to the MAX plan, or to third-party companies building an integration (partner terms unverified). | P1 |
| 5 | Meta Ads | Secondary channel for many contractors | Marketing API. Fragile: attribution windows removed 12 Jan 2026. Data retention limits (13 months for unique and hourly breakdowns). | P1 |
| 6 | Google Business Profile | Calls and direction requests from the map listing | Business Profile Performance API (not checked this run) | P1 |
| 7 | GA4 and Search Console | Expected by agencies that also do SEO | Standard, well documented | P2 |
| 8 | WhatConverts, CallTrackingMetrics | Agencies not on CallRail | Public APIs (not checked) | P2 |
| 9 | JobNimbus (roofing), other trade systems | Next niches on the same engine | WhatConverts already integrates JobNimbus | P2, as an expansion niche |
| 10 | GoHighLevel | Many small agencies run their CRM and forms here | API exists (not checked) | P2 |
| 11 | ServiceTitan | Largest contractors | ServiceTitan has native ad ROI (Marketing Pro). API partner terms unverified. | — at launch |
| 12 | TikTok, LinkedIn, Pinterest, X, Amazon | Rare for contractors | Each is upkeep | — |
4.2 Demanded features, classified
Legend: TS = Table stakes, D = Differentiator, X = Distraction, L = Liability. Priority: P0 = v1 of the recommended wedge, P1 = next, P2 = later, — = don't build.
A. The booked-revenue core (your strongest area, so lean in)
| Feature | Demand signal | Category | Priority | Notes |
|---|---|---|---|---|
| Lead-to-job matching (calls, LSA leads and forms matched to requests, quotes, jobs and invoices by phone, email and time window) | High among agencies whose clients doubt the ads | D | P0 | The engine of the product. WhatConverts does this for its own leads. You do it across sources, on top of whatever call tracking the agency uses. |
| Cost per booked job and revenue per ad dollar, by channel | High. It's the contractor's question. | D | P0 | The report's first line. |
| Unmatched-lead review queue (agency confirms or fixes matches) | Needed for trust | D | P0 | Keeps the numbers honest. Every fix improves the matching rules. |
| "Leads you didn't book" page (missed calls, LSA leads with no job, quotes with no follow-up after 7 days) | High for agencies. Contractors may not enjoy it. | D | P0 | The renewal hook. It shows where revenue was lost after the lead arrived. Let the agency choose whether to include it. |
| Quoted vs sold vs collected revenue | Medium-high | D | P1 | Jobber exposes quotes and invoices. Show all three, with dates. |
| Job-type breakdown (install vs repair vs maintenance) | Medium | D | P1 | Installs are worth 10× repairs. It shows which campaigns bring the valuable jobs. |
| Seasonality comparison (this month vs same month last year) | Medium | TS | P1 | HVAC is seasonal. Month-on-month comparisons mislead. |
| LSA lead quality view (charged vs credited, lead type, service area) | Medium | TS / D | P1 | Google now credits invalid leads automatically (manual disputes ended July 2024), and it doesn't credit "job type not serviced" or "geo not serviced" leads. Showing these helps the agency fix LSA settings. |
| Multi-touch attribution model | Loudly requested by some agencies | L | — | You can't do it accurately with phone matching and platform data. Show matched leads and say so. |
| Revenue forecast ("next month you'll book $X") | Requested | L | — | A wrong forecast is a screenshot that costs the agency the client. |
B. Report and delivery
| Feature | Category | Priority | Notes |
|---|---|---|---|
| White-label PDF, live link and email, custom domain | TS | P0 | Required to be in the category. |
| One template per trade (HVAC, plumbing, roofing, electrical) | D (small) | P0 | Agencies want a finished report, not a canvas. |
| Scheduled sending with agency review step | TS | P0 | The agency approves before the contractor sees it. |
| Contractor-language AI summary, grounded in the report's numbers | TS → D | P0 | Only a differentiator because it talks about jobs and revenue. Cap cost at ~$0.10 per report. |
| Data-gap check before send (sudden zeros, missing fields, expired tokens) | D | P0 | Meta's Jan 2026 change broke reports silently across the category. Catching it is a trust feature. |
| Drag-and-drop dashboard builder | X | — | Looker Studio does it free. |
| Client portal with logins, comments and tasks | X | — | The owner wants a two-minute read. |
| Mobile app | X | — | The report is a link and a PDF. |
| Chat with your data ("Ask AI") | TS | P2 | Incumbents and MCP servers already offer it. |
C. Agency operations
| Feature | Category | Priority | Notes |
|---|---|---|---|
| Self-serve client onboarding (send the contractor an OAuth link for Jobber or Housecall Pro) | TS | P0 | Onboarding friction is the main trial killer. |
| Agency roll-up view (all clients: cost per booked job, at-risk clients) | D | P1 | Shows which clients are likely to churn from the agency. Useful for the agency owner. |
| Team seats and roles | TS | P1 | Account managers per client. |
| Goal tracking per client (target cost per booked job) | TS | P1 | Simple thresholds, not forecasts. |
| Invoicing the agency's clients | X / L | — | Moves you into payments. Not needed. |
| Proposal or pitch reports for prospects | X | P2 | Nice for sales, but a different job. |
D. SEO and local visibility
| Feature | Category | Priority | Notes |
|---|---|---|---|
| Google Business Profile calls and direction requests | TS | P1 | Part of the lead picture. |
| Rank tracking and map-pack grids | X for you | — | BrightLocal, Local Falcon and AgencyAnalytics own it. |
| Review monitoring | X | — | Many review tools, and GoHighLevel includes it. |
| "AI search visibility" tracking | X | — | AgencyAnalytics already sells an AI Tracker add-on. Unproven for contractors. |
E. Trust, data and compliance
| Feature | Category | Priority | Notes |
|---|---|---|---|
| Hashed phone numbers; no homeowner names, addresses or recordings stored | D (trust) | P0 | Easier security answers, less privacy exposure. |
| Connector health page per client | TS | P0 | Expired tokens are the most common support issue in this category (assumed). |
| Audit log of match overrides | TS | P1 | Shows the agency didn't inflate numbers. |
| SOC 2 report | X now | — before $20k MRR | Small agencies rarely ask. Revisit if larger agencies do. |
4.3 Net effect
Of roughly 40 features, about 9 are real differentiators for this wedge, and about 12 belong in v1 (3 connectors plus the matching core and delivery basics). Everything else is table stakes you'll add slowly, or other companies' products you should read from, not rebuild.