first principles
← Channel playbooks

Channel 03 / 13 · Create authority

Original research & benchmarks

Publish something the market can cite, not just another opinion.

Use it when

You have trustworthy aggregate product data, access to practitioners or a defensible benchmark.

Don't confuse it with

Invented statistics, self-serving “surveys” with unclear samples or benchmark charts without methods.

01 / What companies actually did

Harness released a 500-person engineering study

Its 2025 State of Software Delivery surveyed 500 engineering leaders and developers. It published findings and organized a related research event with practitioner speakers.

[H2] Harness — State of Software Delivery (2025) ↗[H3] Harness — Software Delivery research event ↗

Grafana made its observability survey an ongoing asset

Grafana’s 2026 survey had more than 1,300 responses. It published breakdowns about open standards and AI in observability rather than one isolated announcement.

[G1] Grafana — 2026 Observability Survey findings ↗

Harness used measured customer results as proof

The Zeotap case study reports 36× higher deployment frequency and an 83% reduction in deployment effort. This is a company-reported customer result, not a general industry benchmark.

[H4] Harness — Zeotap customer story ↗

02 / How to do it

  1. 01Pick a question practitioners disagree about: cost, reliability, migration time, deployment effort, etc.
  2. 02Select a defensible method: anonymized product data, repeatable benchmark or carefully sampled survey.
  3. 03Document sample size, setup, timeframe, limits and any commercial interest.
  4. 04Publish one flagship report plus 3 findings posts, an explainer, an event and a calculator/checklist.
  5. 05Send different findings to different buyer segments; use the report in sales discovery as evidence.

Things you can publish or run

State of [technical problem] — year / methodology / datasetWhat [N] production workloads reveal about [metric]Benchmarking [A] and [B] under a reproducible workload

03 / What good looks like

Watch this

Qualified conversations, reference links, backlinks, target-account participation and pipeline influenced.

Common trap

A glossy survey is not original research if the method is weak.

04 / Copy the worksheet

COPYABLE WORKSHEET / 03

Original research & benchmarks template

Replace the brackets. Use this as a starting point, not a rigid formula.

RESEARCH-TO-DISTRIBUTION KIT
Question: [one useful industry question]
Evidence: [consented anonymized data | survey | reproducible benchmark]
Sample/method/limits: [...] / [...] / [...]
Three findings: 1) [...] 2) [...] 3) [...]
Publish:
- Full report + methodology
- Three single-finding technical articles
- One engineer-led webinar / discussion
- Five social charts with specific captions
- One buyer-facing checklist or estimator
- Three outreach angles by persona
Follow-up: send relevant finding + practical offer
Measure: citation links, engaged target accounts, opportunities

05 / Original sources

Next channelOutbound & account-based GTM →