Every brand in the benchmark is scored against the same written rules, and every score links to the posts, emails and pages behind it. This page is the full rulebook, so you can check our work or score any brand by hand.
Get your reportEach layer is weighted equally. The score is rungs reached out of 18, as a percentage, rounded. With this table you can score any brand by hand.
Founder voice
Last 30 days
Team advocacy
Last 30 days
Email relationship
24-hour watch
Long-form
Podcast or video 5+ min
Blog
Interactive tools
Blog titles in any of these formats count as written for search.
Take the newsletter or podcast name from the signup form, the signup page title, the sender name or the podcast listing. Strip the company name, articles such as "the", connecting words (for, of, and, by, from), and these generic words.
If any word is left, the channel is named. When the sources give different names, the channel is named if any one of them passes. "Shorts, Tracksuit's fortnightly newsletter" leaves "Shorts", so it is named. "The Tracksuit Weekly" leaves nothing.
Shown above the score. Checked from the top, first match wins.
Scorecard spec v5, 21 Oct 2026.
The rubric says what each rung needs. These are the rules that decide what counts toward it.
Does a founder post their own thinking, week after week?
How many employees post on LinkedIn?
Does the brand run a list with its own identity, and does it say hello?
Does the brand publish podcast episodes or videos of five minutes or more, on a steady rhythm?
Is the blog active, written by people, and written for readers ahead of search engines?
Does the brand give visitors something useful to use, ideally a tool that tells them something about their own business?
Every brand is scored by AI agents working from a written rulebook, the same rules shown on this page. The agent records every item it looked at, whether it counted, the rule it passed or failed, a link to the item and its date.
Before publishing, we tested the rules against Pistachio founder Isaac Peiris's own blind scores. On 7 Oct 2026 he scored nine brands layer by layer and ranked them on gut feel. Two agents then scored the same brands separately, from the rules only. Every disagreement was traced to the rule behind it, and the rule was rewritten.
Calibration brands: beehiiv, Ordinal, Tracksuit, Zapier, Notion, monday.com, Ploy, Linear, DocuSign
94%
of layer scores within one rung of Isaac's blind score
Against a 90% target, across the five layers then in play, 8 Oct 2026. Email, scored on 12 Oct under the 24-hour rule, landed within one rung for all 9 brands.
26 of 27
layers scored identically by two separate agents
The one split was Tracksuit's blog, whose site blocks automated reads.
9 of 9
brands within one place of Isaac's gut ranking
After the rescore on 21 Oct 2026, counting ties as shared places.
A second, independent scorer re-scores a random 10% of brands. Every disagreement is traced to a rule.
A case the rules don't settle goes to Isaac. He rules, and the ruling is written into the spec with its date.
Scores are never changed by hand. If a result looks wrong, the cause is traced to a rule, and the rule is what changes.
Every row behind the report is checked against the scoring file by exact match after import.
Isaac built the list in rounds between 12 and 19 Oct 2026, from a pool of 577 candidate brands.
B2B SaaS and tech companies based in Australia, the UK or the US that had raised a Series A or B in the previous six months, plus Y Combinator alumni at growth stage.
The best-known challengers in GTM tools, HR tech, developer tools, B2B fintech and marketing tools, with no funding window.
For each brand kept, a search for challengers at a similar stage. Challengers only, never incumbents, so Clarify was added alongside Attio and Salesforce was not.
Kept tools sold to a business function and professional or financial-services verticals. Cut physical-industry verticals, media, government, consumer and creator tools.
Isaac cut the list to 109 brands, each his own call.
20 brands added for name recognition across categories, such as 37signals and ActiveCampaign.
6 brands that have sponsored creator content. With one acquired brand removed, the final list is 134.
The list is mainly Australia, the UK and the US. Four exceptions are Isaac's call: Legora (founded in Sweden), Ahrefs (Singapore), Viktor (Warsaw and Munich) and monday.com (Tel Aviv). Calibration, wide-appeal and sponsor brands sit outside the original funding window by design.
Each layer is read from public sources, inside a window counted back from the scoring date.
The founder's LinkedIn posts from the last 30 days, plus podcast guest spots from the last 90.
Every employee's LinkedIn posts from the last 30 days, or a fixed sample at companies above 200 staff.
A signup to the brand's list with one research address per brand, and the inbox read 24 hours later.
The brand's YouTube channel and podcast listings, with dates and lengths.
The blog listing or sitemap, and the last 10 posts with their authors and dates.
One shared pool per brand, built from the sitemap and a fixed set of site searches (calculator, quiz, grader or assessment, benchmark, ROI, "free tools", guide or playbook or report).
The benchmark measures whether a brand shows up in the channels it owns, and how consistently. It does not judge how good the content is.
A score without evidence is not shipped. Each brand's full working file, with every item checked, is kept in Pistachio's records.
Your place against every scored brand on the list. It is recomputed each time a batch is scored, so it moves as the list fills. With 134 of 134 brands scored today, percentiles are against those 134.
Scores are refreshed on request, not on a schedule. A rescore runs the same rules on fresh evidence and replaces the brand's result, and the previous result stays in our records. If your brand is on the list and a layer looks wrong, or out of date, ask us to check it again.
Ask for a rescoreEvery change is Isaac's call, written into the spec with its date. A new ruling makes a new version, and this page always describes the current one.
One field. Your own result if we have scored you, and the full benchmark either way.
Pistachio