Methodology
What we measure, how we measure it, and where the numbers can mislead you. Last updated November 17, 2026.
Data source
We read public job posts on company career pages. Nothing else: no logins, no purchased data, no tracking of visitors and no personal data. This run covers 498 companies, 388 of them funded startups from the Y Combinator directory, and 109 tools across 8 categories.
How a role lands in a category
For each job post we extract the skills and tools it names. A role counts toward a category when its skills include one of the category's tools or keywords, or when its title, department or team contains a keyword. One role can count toward several categories. Tools that match the company's own name are ignored, so a company is never scored for its own product.
The intent score
Each company gets a score from 0 to 100 for every category it is hiring into. It is the sum of six signals, rounded and capped at 100. It estimates how likely the company is to be evaluating tools in that category. It is a prioritization aid, not a prediction.
| Signal | Points | How it is scored |
|---|---|---|
| New roles | 30 | Roles in the category posted in the last 7 days. 30 x min(new roles, 6) / 6, so 6 or more earns the full 30. |
| Open roles | 20 | Open roles that match the category. 20 x min(open roles, 10) / 10, so 10 or more earns the full 20. |
| Share of hiring | 15 | Category roles divided by all the company has open. 15 x min(2 x share, 1), so half of all hiring earns the full 15. |
| Senior roles | 10 | Roles with staff, principal, lead, head, director, VP, vice president, founding, architect or manager in the title. 10 x min(senior roles, 3) / 3. |
| Momentum | 15 | A first-time hire in the category earns 15. Otherwise, if open roles grew since the last snapshot, 15 x min(growth, 5) / 5. No growth or no earlier snapshot earns 0. |
| Size fit | 10 | Team of 30 or fewer x 0.7, up to 500 x 1.0, up to 2,000 x 0.75, above 2,000 x 0.5. Unknown size: startup x 0.7, large company x 0.55. Multiplied by 10. |
Score bands: Hot is 70 and above, Warm is 45 to 69, Watch is below 45.
Definitions
- New role. Posted in the last 7 days, using the posting date on the career page. A role with no date counts as new only if we first saw it in the latest run.
- First-time hire. A company that had no open roles in a category at the previous snapshot and has at least one now. It needs an earlier snapshot to compare against, so it appears from the second run onward.
- Surge. A company with 3 or more new roles in one category within 7 days.
- Competitor mention. A company whose job posts name a tool you track, for example Datadog or Snowflake. Counted from the tools extracted from its posts.
Vendor exclusion
Companies that sell a product in a category are removed from that category. Datadog is not scored as a buyer of observability, and Snowflake is not scored as a buyer of data warehousing. Otherwise a vendor's own hiring would look like buying intent. The list is curated per category and is not exhaustive.
Categories, tools and keywords
Observability & monitoring
You sell to: VP Engineering, Head of Platform, SRE lead.
Tools: Datadog, Splunk, New Relic, Prometheus, Grafana, OpenTelemetry, Dynatrace, Honeycomb, Chronosphere, Sumo Logic, Sentry, PagerDuty, Elasticsearch.
Keywords: SRE, Site Reliability, Observability, Monitoring, Incident Management.
Data stack & warehousing
You sell to: Head of Data, Data Engineering manager.
Tools: Snowflake, Databricks, dbt, Airflow, BigQuery, Redshift, Kafka, Spark, ClickHouse, Fivetran, Airbyte, Dagster, Confluent, Looker, Tableau, Great Expectations.
Keywords: Data Engineering, Data Modeling, ETL, Data Warehouse, Data Pipelines, Analytics Engineering.
AI, LLM & ML infrastructure
You sell to: Head of AI/ML, VP Engineering, CTO.
Tools: OpenAI, Anthropic, LangChain, LlamaIndex, Pinecone, Weaviate, Milvus, Qdrant, pgvector, Hugging Face, Vertex AI, SageMaker, Amazon Bedrock, PyTorch, MLflow, Kubeflow, Weights & Biases, vLLM.
Keywords: LLM, RAG, Generative AI, Machine Learning, MLOps, Deep Learning.
Security & compliance
You sell to: CISO, Head of Security, VP Engineering.
Tools: Okta, Auth0, CrowdStrike, SentinelOne, Zscaler, Palo Alto Networks, Snyk, Vanta, Drata, HashiCorp Vault, Wiz, Jamf, Intune.
Keywords: SOC 2, IAM, Compliance, Security Engineering, Application Security, Zero Trust, SIEM, Penetration Testing, GDPR, HIPAA, ISO 27001, FedRAMP.
Cloud & infrastructure
You sell to: VP Infrastructure, Head of Platform, CTO.
Tools: AWS, GCP, Azure, Kubernetes, Terraform, Pulumi, Crossplane, Istio, Docker, Cloudflare, Vercel, Netlify, Heroku, Fly.io, DigitalOcean, Oracle Cloud.
Keywords: Infrastructure as Code, Platform Engineering, Cloud Infrastructure, DevOps.
Developer tools & CI/CD
You sell to: VP Engineering, Head of Developer Experience.
Tools: GitHub, GitLab, Jenkins, ArgoCD, LaunchDarkly, Statsig, Retool, Jira, Temporal.io.
Keywords: CI/CD, Developer Experience, Developer Productivity, Build Systems, Release Engineering.
Databases & storage
You sell to: VP Engineering, Head of Platform, Database lead.
Tools: PostgreSQL, MongoDB, Redis, MySQL, Supabase, DynamoDB, Cassandra, CockroachDB, Elasticsearch, Algolia.
Keywords: Database, Distributed Systems, Storage.
Go-to-market & product analytics
You sell to: RevOps lead, Head of Growth, VP Marketing.
Tools: Salesforce, HubSpot, Marketo, Segment.io, Mixpanel, Amplitude, PostHog, Heap Analytics, Zendesk, Intercom, Gong.io, Outreach.io, Twilio, NetSuite, ServiceNow.
Keywords: Growth Engineering, Revenue Operations, Marketing Technology, Experimentation, A/B Testing.
Refresh schedule
Career pages are read daily. Scores and account pages update with each run. The weekly report arrives every Monday, and Growth alerts go out the day a first-time hire, surge or competitor mention is detected.
Limitations
- A job post is a signal, not a purchase order. Many roles never lead to a tool purchase.
- Companies set their own posting dates, and some refresh old roles, which can make them look new.
- Career pages can lag behind reality, and some roles are never posted publicly.
- Skill extraction can miss a tool or match one loosely. Open the linked job post to confirm.
- Team size comes from the Y Combinator directory, so it is unknown for large companies and can be out of date.
- Until two snapshots exist, momentum and first-time hires are zero.
Spot an error? Tell us at brett@matrixx.technology.