The Title-to-Level Reference for Cross-Company Talent Maps

This is a lookup document for placing a raw job title at a comparable seniority level so a cross-company talent map lines up. It is written for strategy and research teams, talent-intelligence analysts, and operators sizing a market, who need the same title to mean the same thing whether it came from a 30-person startup or a bulge-bracket bank. Jump to the row you need: the IC ladder, the banking ladder, ownership titles, or company-size distortion. Each entry says what the title proves and how it lies.

The core problem is that a job title is a string chosen by an employer for its own reasons, not a rank stamped by an outside authority. No public taxonomy fixes this for you. So the work is to resolve every string to one normalized seniority level plus one function, and to know in advance which title patterns systematically inflate or deflate.

Why no public taxonomy hands you a seniority ladder

There is no government taxonomy that encodes corporate management seniority. The two public structures that exist measure something adjacent, so you cannot lift a ladder out of them.

O*NET uses Job Zones, which group occupations by the experience, education, and training they need. The original system had five zones, and it is now consolidating to four by merging Job Zones 1 and 2 into a single Job Zone 1-2 in response to declining low-skilled work. That measures preparation, not rank. The Standard Occupational Classification is a separate and larger structure: the 2018 SOC has 23 major groups, 98 minor groups, 459 broad occupations, and 867 detailed occupations. Critically, the SOC is task-based and does not differentiate occupations by education, certification, or seniority. It tells you what work is done, not how senior the person doing it is.

So the seniority layer is practitioner-built. The working consensus is five IC rungs plus a four-rung management track. You build that ladder once, then map every incoming title onto it. The SOC's 23 major groups remain the most defensible public backbone for the function axis, because they are stable, documented, and not tied to any one company's naming.

The dual-axis model: function plus seniority

Resolve every title into exactly one function and exactly one seniority level, decomposed from a four-part parse. This is the shape of a clean row, and it is what makes titles from different companies comparable.

Research on professional-network data decomposes a title into a primary function that is compulsory, a domain that is optional, a position that indicates seniority level, and a secondary function that is optional. Practitioner pipelines mirror this with a sub-function or specialty, such as DevOps or Enterprise Sales, plus a seniority level treated as the organizational authority tier. The point of the parse is that "Senior Staff DevOps Engineer, EMEA" is not one atom. It is a function (engineering), a domain (DevOps), a seniority-position (Senior Staff), a secondary hint, and a region you discard.

Keep the two axes physically separate in your table. The most common data-hygiene failure is stripping a word that carries the level, so before you clean anything, copy the seniority qualifiers into their own column. Dropping "Principal" or "Lead" to isolate the function silently deletes the rank you were trying to measure.

The IC ladder: where Staff out-ranks a Manager

Senior is the last rung the two tracks share. After Senior, the individual-contributor track and the management track run parallel as peers, so a Staff or Principal IC is not automatically below a person with "Manager" in their title.

At most major tech companies the IC track after Senior goes Staff, Senior Staff, Principal, Distinguished. The management track goes Engineering Manager, Senior EM, Director, VP of Engineering. These are peer bands, not a stack. The documented anchors are precise: at Google, Staff is L6 and Engineering Manager is also L6; at Microsoft, Principal at band 64 shares that band with an EM; at Meta, E6 and M1 are peers. A Staff IC out-ranks a first-line "Engineering Manager" only when that manager is an M1 running a single team while the IC sits at Senior Staff or Principal. A bare "Manager" of one team is roughly Staff-equivalent, and Principal or Distinguished sits above it.

Standardization is a function of headcount. Firms above 500 employees run numbered ladders, Google L3 to L10 and Meta E3 to E9, which makes their titles reliable enough to anchor the rest of your map. The same title is trustworthy at 5,000 people and near-meaningless at 30.

Staff versus Principal is not a clean swap

The two post-Senior IC titles do not level the same way. In Refolk's index, "Principal Engineer" appears on about twice as many US profiles as "Staff Engineer," and that ratio tells you something about how the words are used.

Source: count columns from Refolk's index; the multiple is derived by division. Top employers for Staff include Stripe and OpenAI; for Principal, Apple, NVIDIA, Amazon, and Cloudflare. Because "Principal" is applied roughly twice as often, it spans a wider band of real seniority: some firms use it as the first post-Senior rung, others as near-Distinguished. So Principal needs company context to level, while Staff is tighter. Do not treat them as interchangeable rungs.

The banking ladder: the biggest inflation trap

A banking VP is not an executive. In financial services and banking, titles inflate heavily, and a "VP" there often equals a Manager or Senior Manager elsewhere. This is not an edge case; it is a volume phenomenon that will skew any finance-heavy map.

The banking ladder runs Analyst, Associate, VP, ED or SVP, MD, with Partner at the top. The titles that truly signify senior leadership in finance are Executive Director, Managing Director, and Partner. An ED or MD is the functional equivalent of a VP or SVP in a standard corporation. Banks grant the VP title as a client-facing officer badge, which is why a VP arrives roughly six to eight years into a career rather than at the executive tier.

The scale of the distortion is the reason this deserves its own row. In Refolk's index, "Vice President" is more than eighteen times as common in finance as in software.

Source: both count columns from Refolk's index; the ratio is derived by division. Top VP employer in finance: Goldman Sachs. Any map that treats "VP" uniformly systematically over-ranks that 71,682-profile slice.

Here is the normalized resolution for finance titles.

Source: banking ladder and equivalences from public career-path write-ups. Note that "Analyst" collides across industries: a banking Analyst is an entry IC, while an "Analyst" in a data function is something else entirely. Use the employer's industry and the person's adjacent skills to break the tie.

Company size distorts scope more than any string rule

Company headcount shifts real scope substantially, and for sales it is a bigger distortion than industry. The same "VP of Sales" title covers a person running a team of five who still sells half their time, and a person sitting three layers above every rep who never touches a deal.

Director-to-VP inflation is accelerating at small companies. 18% of companies with fewer than 50 employees now post VP Sales roles whose scope matches a Director job at a larger company. Related drift shows up in early-career titling too: the use of "Lead" in job descriptions for early-career roles tripled between 2019 and 2023. So bucket every employer by headcount and discount inflation accordingly.

Source: headcount bands and inflation figures from public sales-leadership and job-title-inflation research. The practical read: at a 30-person firm a VP is a hypothesis, at a 5,000-person firm a VP is close to a fact.

Refolk resolves the title against the fields that actually settle the level. When you ask Refolk for people by role, stage, headcount, and reporting line in plain English, the size band and scope come back attached, so you are not leveling a bare string.

Ownership titles carry no fixed scope

Founder, Owner, and CEO are unbounded titles that prove nothing about scale on their own. A solo operator and the chief executive of a 5,000-person company can both wear "CEO," so these strings must be leveled entirely by headcount and reporting structure.

Refolk's index confirms the looseness directly: "Owner," "CEO," and "Founder" all appear inside a VP-titled financial-services population, which means the strings float free of any consistent rung. Never map an ownership title to executive scope without checking headcount and reports. A "Founder" at a two-person company is not a peer of a "Founder" who now runs a large org.

An ownership title is a claim about who holds the equity, not a measure of how far the authority reaches.

How this goes wrong: failure modes and false positives

Most bad talent maps fail on a handful of repeatable errors, and every one of them has a cheap check. This is the section to keep open while you work.

• Reading banking VP as executive. False positive: a 26-year-old Goldman "Vice President" mapped to C-suite-adjacent. Check: if the employer industry is finance, downgrade to Manager or Senior Manager, and confirm via tenure, since a banking VP typically arrives six to eight years in.
• Trusting startup VP and Head-of titles. False positive: a "VP of Sales" at a 30-person firm scored as an enterprise VP. Check: look at team size, budget authority, and board presentation. If all three are "no" or "eventually," it is a Director role wearing a VP title.
• Assuming management out-ranks IC. False positive: an Engineering Manager scored above a Staff or Principal IC. Check: the tracks are peers at L6, band 64, and E6. Level by band, not by track.
• Stripping seniority words that carry function. False positive: dropping "Principal" loses the level, and dropping "Lead" mis-buckets an inflated early-career role. Check: capture seniority in a separate column before you strip anything.
• Function collision on "Analyst." False positive: a banking Analyst, an entry IC, mapped into the data-analytics function. Check: use industry plus adjacent skills to disambiguate, because "Analyst" in finance and in healthcare need different treatment.
• Ownership-title overweighting. False positive: a solo "Founder" or "Owner" scored as a CEO at scale. Check: read headcount and reports; ownership strings carry no scope by themselves.
• Stale maps. False positive: last year's mapping still trusted after titles have drifted. Check: re-score a sampled cohort every year.
• Over-trusting title-only confidence. False positive: high confidence from a clean string that is actually inflated. Check: require at least one non-title field, reports-to or team size, before you lock the level.

The procedure: resolving one title to one level

Run every title through the same eight steps. The string gives you a first guess; the overrides and scope fields give you the final level and a confidence score.

The scope check in step seven is where the accuracy lives. The reliable read is always the underlying scope: how many people report up, how much revenue or budget the role owns, and where it sits in the reporting line. This is not just a human heuristic. Automated seniority-normalization systems identify profiles of similar seniority across organizations using models trained on salaries, job functions, and working experience. Both the practitioners and the machines converge on the same non-title signals, so a pipeline that stops at the string is leaving documented accuracy on the table.

raw_title | normalized_function | title_only_level | final_level | override_type | reason
VP, Investment Banking | Finance | VP | Senior Manager | industry | banking VP officer badge; 7 yrs tenure
VP of Sales | Sales | VP | Director | company_size | 34-employer headcount; team of 4
Staff Engineer | Engineering | Staff | Staff | confirmed | reports to Director of Eng, no downlines
Founder & CEO | Executive | CEO | Manager | ownership | 3-person company, no reports

Before you trust the map

Run this checklist before you publish a leveled map or hand it to a decision-maker. Each item catches a specific failure from the section above.

Keeping the map current

Titles inflate quietly, so a mapping you trusted last year drifts without any single event to warn you. Re-score a sample of executive roles each year and confirm the title still matches the evaluated level, then date the audit.

Two mechanisms will move under you. First, the words themselves drift: "Lead" tripling in early-career descriptions over four years is the kind of shift that turns a reliable rule into a false positive. Watch the high-volume ambiguous strings, VP, Head-of, Lead, Principal, and re-check their real scope distribution rather than assuming last year's mapping holds. Second, the public backbone moves on a slower clock: the SOC 2018 will be reviewed for possible revision, so if you anchor your function axis on it, note the version and plan to re-map when it changes. The mapping is never finished. It is a living table with a date on it and a cohort you re-score, and that is the difference between a talent map you can defend and one that quietly lies.