Broader product bets

This page deliberately looks past the current product descriptions. It treats the portfolio as a set of reusable observation systems, data assets, and evidence-backed workflows that can create products beyond the first app each repo describes.

The point is not to chase every idea. The point is to see where the same technology and data can compound.

Portfolio assets

Asset	Current sources	What it can become	Why it matters
Domain and website observation ledger	SiteIntel, Trackpal, DataStitcher public audit ideas	A time-series record of what changed across websites, domains, pages, tags, schemas, DNS, TLS, robots, sitemaps, and public risk/reference signals.	Domain change is commercially useful for agencies, sales teams, investors, compliance teams, and operators.
Public dataset and schema ledger	BigQuery Atlas, DataStitcher, Meridian BigQuery packages	A monitored catalogue of public datasets, schemas, freshness, breaking changes, and derived dashboards.	Public data users often do not know when source data changed, disappeared, or became unsafe to trust.
Evidence and provenance model	Bollocks Radar, SiteIntel, BigQuery Atlas, DataStitcher	A reusable way to represent a claim, finding, candidate, source, confidence, caveat, and evidence trail.	This is the difference between a useful intelligence product and a confident but brittle summary tool.
Business metric snapshots	Trackpal, DataStitcher	A normalised history of spend, performance, freshness, alerts, and template-driven business metrics.	Small teams buy products that tell them what changed and what needs attention, not generic dashboards.
Ingestion run history	DataStitcher, SiteIntel, BigQuery Atlas, Bollocks Radar	A reusable operational graph of sources, runs, retries, artefacts, failures, costs, and freshness.	This is product infrastructure: every future data product needs it.
AI-agent delivery telemetry	Codex Sessions Analyser, Meridian Hub, Product Compass	A private delivery ledger showing cost, tool usage, PR cycle time, failure modes, and context quality.	AI-assisted development needs measurement or it becomes vibes, expense, and hidden rework.

Compounding principles

Fact: The current portfolio already repeats the same core moves: collect evidence, normalise it, preserve provenance, run bounded workers, store raw artefacts separately, generate reports, and expose operator/admin status.

Inference: The larger opportunity is not one more app. It is a set of low-cost intelligence products built from the same observation and evidence machinery.

Opportunity: Build ledgers before building dashboards. A durable event/history model for domains, datasets, provider syncs, claims, and AI work will create more optionality than another bespoke UI.

Opportunity: Use domain, source, time, and workspace as the four primary join keys:

• Domain joins websites, businesses, competitors, public records, agency clients, and source references.
• Source joins facts, caveats, freshness, legal wording, and trust boundaries.
• Time turns one-off checks into monitoring, alerts, trends, and change intelligence.
• Workspace keeps private customer/provider data separate from public observations.

Product ideas beyond the current apps

Internet Change Ledger

Opportunity: A cross-domain change intelligence product that records what changed across websites, domains, public datasets, ranking feeds, public safety/reference feeds, and source pages.

Possible buyers:

• Agencies monitoring client and competitor portfolios.
• Investors or acquirers watching company/product movement.
• B2B sales teams looking for trigger events.
• Operators watching supplier, partner, or franchise sites.
• Researchers tracking public web and dataset drift.

First useful wedge:

• Track 50 to 500 domains.
• Show material weekly changes: homepage title/meta, important page status, robots/sitemap changes, schema changes, DNS/TLS expiry, visible platform changes, and selected public signal changes.
• Send a weekly "what changed and why it may matter" digest with source links.

Existing leverage:

• SiteIntel domain tasks, SiteIntel domain registry, bulk source ideas, raw artefact boundary, and evidence-backed report pattern.
• DataStitcher source/run model.
• BigQuery Atlas candidate/evidence mindset.

Why it is broader:

• This is not just "website monitoring". It is a structured history of public internet change that can power many vertical tools.

Cautions:

• Keep it focused on public business/site signals.
• Avoid creepy people tracking.
• Avoid vulnerability-scanner positioning unless the product is explicitly built and reviewed for that risk.

Agency Intelligence OS

Opportunity: A higher-value agency product that combines client spend health, website health, market/context changes, and generated account-review narratives.

Possible buyers:

• Small and mid-sized marketing agencies.
• Freelance growth consultants.
• Multi-site operators who manage digital channels across locations.

First useful wedge:

• Monthly client review pack for an agency account manager.
• Inputs: Trackpal spend/budget pacing, SiteIntel website change/health checks, simple competitor/domain observations, and optional local/business context from DataStitcher.
• Output: a client-ready narrative: "what changed, what needs action, where budget is pacing badly, what evidence supports it."

Existing leverage:

• Trackpal templates, metric catalogue, provider/sync health, and agency orientation.
• SiteIntel domain checks and findings.
• DataStitcher semantic registry and public data products.
• Product Compass evidence/report rules.

Why it is broader:

• Trackpal is not just dashboards; it can become the operating layer for account reviews and client communication.

Cautions:

• Do not become a generic dashboard builder.
• Do not over-automate advice until evidence and caveats are strong.
• Keep private client data isolated from public portfolio data.

Dataset Reliability Monitor

Opportunity: Monitor public datasets, BigQuery tables, CKAN sources, government feeds, and derived data products for freshness, schema drift, missing files, changed definitions, and broken downstream dashboards.

Possible buyers:

• Data engineers using public datasets.
• Analysts and researchers.
• Newsrooms and civic data teams.
• SaaS teams that depend on public reference data.

First useful wedge:

• Watch a known list of BigQuery public datasets and public CKAN/API sources.
• Publish schema/freshness changes and a weekly digest.
• Expose a small API for dataset status, last-seen schema, and breaking-change notes.

Existing leverage:

• BigQuery Atlas catalogue model and cautious validation.
• DataStitcher monitored ingestion and semantic registry.
• D1/R2 compact-plus-heavy-payload pattern.
• Generated docs for source caveats and status.

Why it is broader:

• BigQuery Atlas can be more than discovery. It can become the "is this public data still usable?" layer.

Cautions:

• Bound BigQuery costs.
• Treat freshness and reliability as evidence, not as source quality judgement unless criteria are explicit.

Evidence-Backed Briefing Engine

Opportunity: A briefing system that turns public observations, internal metrics, and source evidence into short, cited, decision-ready memos.

Possible buyers:

• Agencies preparing client reviews.
• Founders researching competitors or markets.
• Consultants preparing prospect briefs.
• Operators assessing suppliers, partners, or geographic expansion.

First useful wedge:

• "Brief this domain or business" with a deterministic evidence pack first, then optional AI narrative.
• Include citations, freshness, caveats, and "unknown" states.
• Use saved briefing templates for agency review, competitor scan, public-data scan, and claim/source context.

Existing leverage:

• SiteIntel report story.
• Bollocks Radar source attribution discipline.
• DataStitcher generated docs and semantic registry.
• Product Compass AI context pattern.

Why it is broader:

• The product is not the crawler or dashboard. The product is repeatable, evidence-backed synthesis.

Cautions:

• AI should summarise evidence, not invent evidence.
• Put caveats beside claims, not in a footnote nobody reads.
• Keep raw/private data out of public or shareable reports by default.

Trust and Claim Context API

Opportunity: A narrow API/widget layer that provides source-attributed context for claims, topics, publishers, and public references.

Possible buyers:

• Publishers and newsletters.
• Communications teams.
• Educators and civic organisations.
• Moderation/research tools that need source context, not automated truth labels.

First useful wedge:

• Search a claim or topic.
• Return matching fact-check reviews, source links, dates, publishers, verdict wording, and a clear "no known match" response.
• Add domain/source context later.

Existing leverage:

• Bollocks Radar Google Fact Check ingestion, verdict normalisation, claim clustering, and source metadata.
• SiteIntel domain/source signals.
• Evidence/provenance model from Product Compass.

Why it is broader:

• Bollocks Radar can be more valuable as context infrastructure than as a standalone consumer search app.

Cautions:

• Do not build a "truth oracle".
• Do not infer falsehood from absence of evidence.
• Legal and reputational wording needs extra care.

Provider Health Watchtower

Opportunity: A product or reusable admin layer that watches connected providers, OAuth/scopes, secret/config status, sync freshness, quota warnings, and stale data across many small SaaS/admin systems.

Possible buyers:

• Internal Meridian products first.
• Small SaaS operators with multiple provider integrations.
• Agencies managing many client provider accounts.

First useful wedge:

• One operator screen listing providers, expected secrets, last sync, latest failure, stale data, scope/config blockers, and manual retry actions.

Existing leverage:

• Trackpal provider registry and sync health.
• Meridian Hub operational model.
• Product Compass provider/source map.
• Repeated secret-status/admin patterns.

Why it is broader:

• The repeated pain is not just "connect Google Ads". It is "know which integration broke before customers find out".

Cautions:

• Never expose secret values.
• Treat provider-specific scopes and approval state as source-owned facts.
• Start as internal infrastructure before trying to sell it.

Public-Data Product Studio

Opportunity: A productised workflow for turning public datasets into governed mini-products: ingestion, registry, caveats, marts, generated docs, dashboard templates, and API/export surfaces.

Possible buyers:

• Consultants building niche data products.
• Civic organisations.
• Internal use for Meridian product experiments.
• Analysts who repeatedly turn messy public datasets into explainable outputs.

First useful wedge:

• A CLI/admin workflow that takes one source config and produces a monitored ingestion, source docs, freshness page, and simple dashboard/API.

Existing leverage:

• DataStitcher platform/data-product separation.
• DataStitcher registry and monitored ingestion.
• @meridian-stack/works-docs generated documentation.
• Trackpal template-driven dashboard thinking.

Why it is broader:

• DataStitcher can become a factory for multiple tools, not only a single public data website.

Cautions:

• Avoid "generic ETL platform" positioning.
• Choose opinionated vertical outputs and strict source caveats.

Local Market and Competitor Observatory

Opportunity: A market-intelligence product that tracks businesses, websites, categories, locations, public data, and competitor web changes for a region or niche.

Possible buyers:

• Agencies pitching local businesses.
• Franchisors and multi-location operators.
• Economic development organisations.
• Niche B2B consultants.

First useful wedge:

• "Show me the public web and business-data picture for cafes/plumbers/dentists in this region."
• Include website presence, site health, schema, public business records where permitted, visible platform hints, and change history.

Existing leverage:

• SiteIntel domain and website analysis.
• DataStitcher official/public sources.
• BigQuery Atlas public dataset discovery.
• Evidence-backed briefing engine.

Why it is broader:

• It combines public data and web observation into a defensible niche-intelligence product.

Cautions:

• Avoid lead-spam tooling.
• Do not publish sensitive or name-bearing records without explicit publication policy.
• Keep official-source caveats visible.

AI Delivery Ledger

Opportunity: A private developer/team product that shows what AI-assisted development actually costs, where it gets stuck, which tools fail, which prompts work, and what code/PR outcomes result.

Possible buyers:

• Scott/Meridian first.
• Small engineering teams using Codex/GitHub Copilot heavily.
• Consultants billing or estimating AI-assisted work.

First useful wedge:

• Local report that links sessions, tool calls, cost estimates, PRs, elapsed time, failed commands, and final outcomes.
• Add redacted team exports only after privacy rules are solid.

Existing leverage:

• Codex Sessions Analyser JSONL parsing, SQLite cache, cost estimation, local dashboard, and redacted export concept.
• Meridian Hub repo analysis reports.
• Product Compass context and project docs.

Why it is broader:

• Codex Sessions Analyser is not only a cost tool. It can become delivery observability for AI-assisted engineering.

Cautions:

• Hosted raw-session SaaS is a bad default.
• Keep local-first and redact aggressively.
• The value is outcome quality and bottleneck diagnosis, not token charts alone.

Best near-term bets

Bet	Why it deserves attention	First experiment	Expected payoff
Agency Intelligence OS	Clear buyer, combines Trackpal and SiteIntel, creates higher-value narratives than dashboards.	Generate one monthly client-review pack from demo Trackpal metrics plus SiteIntel checks.	Commercial SaaS wedge with low extra infrastructure cost.
Internet Change Ledger	Turns one-off domain checks into a durable data asset.	Track 100 public domains for weekly material changes and write digest output.	Reusable data layer for monitoring, sales triggers, research, and agency products.
Dataset Reliability Monitor	Strong fit for BigQuery Atlas and DataStitcher, underserved niche.	Watch 20 public BigQuery datasets and 10 CKAN sources for schema/freshness changes.	Distinct data product with API/digest monetisation potential.
Evidence-Backed Briefing Engine	Converts technical collection into user-facing value.	Build one briefing template for domain/business review with citations and caveats.	Makes every data product easier to sell and understand.
AI Delivery Ledger	Internal value is immediate and can improve every other build.	Link Codex session metrics to branch/PR/check outcomes locally.	Better AI workflow, cost control, and reusable internal tooling.

Ideas to avoid or defer

Idea	Why to be careful
Generic dashboard builder	Trackpal and DataStitcher should use dashboards as delivery surfaces, not compete with BI tools.
Generic public-data lake	Public data only becomes valuable when tied to a clear job, buyer, caveat model, and freshness promise.
Broad vulnerability scanner	SiteIntel should not drift into security claims without specialist review, legal wording, and abuse controls.
Truth-scoring engine	Bollocks Radar should provide source context and known fact-checks, not pretend to determine truth.
Hosted raw AI-session analytics	Developer sessions can contain private code, secrets, and client context. Local-first is the sane default.
People or personal surveillance products	Domain/business intelligence must stay on the right side of public, commercial, and ethical boundaries.

Build sequence

1. Define shared event shapes: domain observation, dataset observation, provider sync observation, claim/source observation, and agent-work observation.
2. Build one generated "evidence card" component that can display source, timestamp, confidence, caveat, and next action across products.
3. Run one low-cost weekly ledger experiment for domains and datasets.
4. Generate one agency-style briefing from existing demo data.
5. Add capability/source metadata conventions so Product Compass can discover these assets from repos automatically.
6. Only then decide which idea deserves productisation.

Sources consulted

• docs/opportunities/product-opportunity-catalogue.md
• docs/opportunities/strategic-themes.md
• docs/opportunities/saas-candidates.md
• docs/capabilities/capability-map.md
• docs/reusable-components/extraction-candidates.md
• docs/integrations/provider-and-data-source-map.md
• docs/projects/portfolio-overview.md