Job Titles to Seniority & Department - Classify Leads in Bulk

Turns messy job titles like "Sr. Mktg Mgr" or "Co-Founder & CEO" into a seniority level (Founder/Owner, C-level, Partner, VP, Director, Manager, Senior IC, IC, Intern/Entry), a department and a decision-maker flag. Made for sales and lead-gen teams who have thousands of titles from LinkedIn or a leads finder and need to filter to the people who can buy.

Rules only, no AI model, no network: the same title always gets the same answer, and it costs a fraction of a cent.

Quick start (3 copy-paste examples)

1. See it work (empty input): run with {}. It classifies 8 sample titles in English, French, Spanish and German (free).

2. A list of titles:

{ "titles": ["Sr. Mktg Mgr", "Co-Founder & CEO", "Directeur Commercial", "Leiter Entwicklung", "Gerente de Ventas"] }

3. A whole lead dataset (pick the dataset, name the column):

{ "datasetId": "YOUR_DATASET_ID", "titleField": "jobTitle", "keepOriginalFields": true }

Use it with (pipelines)

• harvestapi LinkedIn profile / employee scrapers -> this Actor. Set titleField to position or headline; if a row lacks the field, jobTitle, position, currentPosition[0].position and headline are tried.
• Any leads-finder or Google Maps + contact enrichment output -> this Actor -> filter on isDecisionMaker.
• Per-call use from an AI agent: run the Actor with a short titles list and read the dataset.

Input guide

Field	What it does
Job titles	One title per line. Ignored when a dataset is picked.
Lead dataset	Dataset with a title column (picker).
Title column name	Field holding the title, default title.
Keep the original columns	Copies the original row next to the new columns.
Maximum titles	Safety cap, default 10,000.

Output

Field	Meaning
inputTitle, normalizedTitle	What you gave and the cleaned text the rules saw.
seniority, seniorityRank	Founder/Owner (1), C-level (2), Partner (3), VP (4), Director (5), Manager (6), Senior IC (7), IC (8), Intern/Entry (9), Unknown (10).
department, departments	Executive, Sales, Marketing, Engineering, Product, Design, IT, Data, Finance, HR, Operations, Legal, Customer Success, Procurement, Healthcare, Education, Research, Consulting, Admin or Other.
isDecisionMaker	True for Founder/Owner, C-level, Partner, VP and Director.
buyingRole	decision-maker, influencer (managers), practitioner or unknown.
confidence	high / medium / low.
language	Detected title language (en, fr, es, de).
isFormer	True for "Former CEO" style titles (then not a decision-maker).
matchedRules	Which rules fired, for debugging.

Sample row:

{ "type": "title", "inputTitle": "Co-Founder & CEO", "normalizedTitle": "co founder & ceo", "seniority": "Founder/Owner",
  "seniorityRank": 1, "department": "Executive", "departments": ["Executive"], "isDecisionMaker": true,
  "buyingRole": "decision-maker", "confidence": "high", "language": "en", "isFormer": false, "matchedRules": ["founder", "c-level"] }

How it decides

1. The title is cleaned (accents, punctuation, "at Company" removed) and abbreviations are expanded (Sr, Mgr, Mktg, Dir, Eng, Ops and others).
2. It is split on &, ,, |, "and/et/y/und"; each part is classified by an ordered rule dictionary and the highest seniority wins ("Co-Founder & CEO" is Founder/Owner).
3. Special cases come first: interns and trainees, assistants ("Assistant to the CEO" is an IC), "Assistant Vice President", "Product Owner", "Chief of Staff", "Vice President" (never read as President).
4. Department comes from a keyword dictionary in priority order, per part; the first department found in reading order is the main one.

Languages: English, French, Spanish and German (including German compounds such as Vertriebsleiter, Teamleiter, Marketingmanager, Finanzvorstand).

Measured accuracy (hand-labeled test set written by the author)

Run node test/accuracy.js and node test/accuracy.js --set=heldout_titles.tsv to reproduce.

• Set 1, 195 titles (written while building the rules, so optimistic): seniority 100%, department 100%, decision-maker 100% after fixing three bugs it exposed (first run: department 98.5%).
• Set 2, 60 titles written after the rules were frozen (the fairer number): first run seniority 100%, department 93.3%, decision-maker 100%; four department misses were fixed and the set is now 100%, which is tuned, not independent.
• Both sets were labeled by the same person who wrote the rules, with the labeling conventions in this README (for example any "... Manager" is Manager, "Head of ..." is Director, "General Manager" is Director). On real LinkedIn data with unusual titles, expect lower accuracy than these numbers; it has not been measured on a third-party labeled set.

Price

Pay per event: $0.0003 per title classified. Empty or placeholder titles ("N/A", "-") and the free demo are never charged. Example: 10,000 titles = 10,000 x $0.0003 = $3.00.

FAQ

Why rules and not AI? Same answer every time, nothing sent anywhere, and fast and cheap enough for 100,000 titles.

Is a Manager a decision-maker? The flag is true from Director up. Managers are marked influencer in buyingRole, since budget authority varies by company.

Honest limits

• Only EN, FR, ES and DE vocabulary. Other languages fall back to "IC" or "Unknown" with low confidence.
• Titles are ambiguous: "Principal", "Associate", "Lead" and "Partner" mean different things in different industries; the rules pick the most common reading.
• "Manager" always maps to Manager, even where it is an individual-contributor title (some "Product Manager" roles).
• A part after "|" such as a tagline or company name can add a wrong department; the seniority is unaffected.
• Company size is not considered: a CEO of a 2-person firm and of a bank get the same level.