# Methodology: Do sales reps get promoted faster by changing companies?

By Cristian Frunze, Founder of Ken · Oct 26, 2026 · [Ken Research](/research) No. 14 · Sales teams, Hiring and org design

The full methodology behind [Do sales reps get promoted faster by changing companies?](/research/stay-or-jump-sales-promotions.md): the data, every method, every number and every limitation. The same text as the [methodology PDF](/research/stay-or-jump-sales-promotions.pdf).

Do sales reps who switch companies reach the next rung sooner than reps promoted internally?

The hypothesis and plan were pre-registered on 2026-10-16, before any data was pulled.

## 1. Data

**Ken internal data.** Ken's dataset of public professional profiles, work histories and company records. The people data covers people whose current country is the US, UK, Canada, Australia, Ireland or New Zealand. Each profile lists past and current roles with title, company, and start and end month. Profiles were last refreshed between 2025-11 and 2026-04.

**Sales careers extract.** We reused Ken Research's shared sales careers extract, built for the Ken Research study *Who are GTM engineers, and where do they come from?* on 2026-10-15: every person with at least one sales or marketing title, and all of that person's roles, with each title tagged with a function and a level by an ordered set of title rules. It holds 23,543,766 people and 136,156,211 roles. We changed nothing in it.

**Company records.** Employee count, size band and industry for the companies in the sample. 82,899 company ids were covered by records that Ken's earlier study of sales tenure had already pulled; we pulled records for the other 590,059 ids. In total 493,604 of the 672,958 company ids on sample sales roles matched a record, and 477,828 had an employee count.

**Education.** Highest degree for a random 200,000 people in the sample, used only as a control.

## 2. Sample

- **Ladder.** From the extract's labels: SDR = sales development function at individual-contributor level; AE = account executive function at individual-contributor level; manager, director and VP or C-level = those levels in any sales function (sales development, account executive, general sales, account management, partnerships and business development, customer success, enablement, revenue operations, sales engineering). Founder and intern roles are not rungs.
- **Two cohorts.** The SDR cohort starts at a person's first SDR role and looks for the first later role at AE or higher. The AE cohort starts at a person's first AE role (former SDRs included) and looks for the first later role at manager or higher.
- **Entry window.** The first role must start between 2008-01 and 2023-12. If a person's first SDR or AE role started outside that window they are out; a later role is not used instead.
- **Exclusions.** Roles with no start date, an end before the start, or longer than 20 years were dropped: 5,619,916 of the 136,156,211 roles. Of the 2,302,084 people with an SDR or AE role, people whose first role at the cohort's rung started outside the window (SDR 99,060; AE 693,207), who already held a role at the target rung or higher on or before entry (SDR 52,521; AE 210,049), or whose entry company was unknown (SDR 82; AE 330) were dropped, as were the few (under 50 per cohort) whose promotion company was unknown. Intern roles (intern level or student and intern titles) are skipped on the walk.
- **Final sample.** 256,008 SDR entrants and 1,108,867 AE entrants.
- **Group size.** Every published group counts at least 50 people.

## 3. Methods

**Same company or not.** A company is identified by its company id, or when that is missing, by its name lower-cased with punctuation removed.

**Walking each career.** From the entry role we read every later role in start order:

1. A role at the target rung or higher is the promotion. The path is **internal** if every sales role up to and including it was at the entry company, and a **switch** path otherwise. Switch paths are split into **promoted at the move** (the higher role is at a different company from the previous sales role) and **sideways, then promoted** (the person first moved at the same or a lower rung, then was promoted at the new company).
2. A non-sales or founder role counts as **leaving sales** only if it starts after every sales role seen so far has ended. One that overlaps a sales role is treated as a side role and ignored. Intern roles are ignored. Within a month, sales roles are read before non-sales roles, and a sales role at another company that starts in the promotion month makes the path a switch.
3. Otherwise the person is **censored**: at the end of their last sales role, or at their profile's last update if a sales role is current (capped at 2026-04).

Time is counted in whole months from the entry role's start month to the promotion role's start month.

**Company size at the time of a role.** Ken internal data has no dated company size. For each company we rebuilt a yearly trend from the extract: H(year) = the number of extract people with a role there covering 1 July. Size at the time = today's employee count x H(year of the role) / H(2025). When H(2025) is under 5, today's count is used unchanged. Size bands: 1-10, 11-50, 51-200, 201-500, 501-1,000, 1,001-5,000, 5,001+, and unknown.

**Title inflation.** Pre-registered rule: at companies with 1-50 employees at the time, a director title counts as manager and VP or C-level as director; at 51-200, VP or C-level counts as director. This rule cannot change who counts as manager or higher, so it leaves both headline steps unchanged; we found this after the first run. We therefore added two checks, labelled exploratory because they were not pre-registered: (1) a stricter rule under which manager, director and VP titles at companies of 1-10 employees do not count as a promotion; (2) the comparison restricted to promotions where the destination company is in the same size band as the entry company or larger.

**Estimates.**

- Median and 25th and 75th percentiles of months to promotion among the promoted, internal vs switch. Intervals: 1,000 bootstrap resamples of people (seed 51). Months are whole numbers, so at these sample sizes an interval can be a single value.
- Cumulative incidence (Aalen-Johansen) of internal promotion, switch promotion and leaving sales, as competing risks, by month up to 120.
- Cox model for promotion with a time-varying covariate "now works at a different company than at entry", controlling for entry period, entry company size band at the time, industry (top 12 and other) and country. It runs on a random 300,000 people per cohort (all 256,008 for SDRs), and again on the education sample with highest degree added. A promotion that is itself the move has the switch and the promotion in the same month, so we fit it two ways: "move first" (the switch starts half a month before) and "move with promotion" (the covariate is 0 at that event). The first inflates the ratio mechanically; we treat the second as the conservative estimate.
- Linear regression of log(months + 1) on switch path and the same controls, among the promoted: exp(coefficient) is the ratio of times.

## 4. Results

**Outcomes.** AE cohort: 215,328 promoted, 448,190 left sales, 445,349 censored. SDR cohort: 96,766 promoted, 87,199 left sales, 72,043 censored.

**Routes.**

| Step | Internal | Promoted at the move | Sideways, then promoted | Switch total |
|---|---|---|---|---|
| SDR to AE or higher | 48,717 (50.3%) | 36,433 (37.7%) | 11,616 (12.0%) | 49.7% |
| AE to manager or higher | 59,171 (27.5%) | 121,351 (56.4%) | 34,806 (16.2%) | 72.5% |

**Months to promotion among the promoted** (size-normalized and raw are identical).

| Step | Internal p25 / median / p75 | Switch p25 / median / p75 | Gap in medians (95% interval) |
|---|---|---|---|
| SDR to AE or higher | 9 / 14 / 22 | 15 / 26 / 41 | +12 (12 to 12) |
| AE to manager or higher | 18 / 31 / 50 | 25 / 46 / 73 | +15 (15 to 16) |

By switch type: promoted at the move, medians of 24 (SDR) and 41 (AE) months, gaps of +10 and +10; sideways then promoted, 32 and 61 months, gaps of +18 and +30 (intervals 17 to 18 and 29 to 31).

**Cumulative incidence.**

| Step | Month | Internal | Switch | Left sales |
|---|---|---|---|---|
| SDR to AE or higher | 12 | 8.3% | 3.6% | 7.9% |
| | 36 | 18.1% | 13.8% | 25.9% |
| | 60 | 19.7% | 19.0% | 34.5% |
| | 120 | 20.7% | 23.5% | 42.0% |
| AE to manager or higher | 12 | 0.8% | 1.3% | 5.6% |
| | 36 | 3.3% | 5.8% | 22.7% |
| | 60 | 4.8% | 10.2% | 34.4% |
| | 120 | 6.1% | 17.0% | 46.9% |

**Destination size.** Share of promotions to a company in a smaller size band than the entry company: AE step 48.1% of switch vs 2.5% of internal; SDR step 42.1% vs 3.0%.

**Title-inflation checks (exploratory).** Destination the same size band or larger: AE step 31 vs 49 months, gap +18 (17 to 18), switch share 48.9%; SDR step 14 vs 26, gap +12 (11 to 12). Strict rule: AE step 31 vs 47 months, gap +16 (15 to 17), switch share 72.2%.

**Breakdowns.** Gap in medians, AE step, by entry period: 2008-2011 +18, 2012-2015 +17, 2016-2019 +13, 2020-2023 +5; SDR step: +18, +19, +14, +9. By entry company size band: AE +14 to +17, SDR +11 to +14. By country: AE +13 to +16, SDR +10 to +12.

**Adjusted models.**

| Step | Sample | Hazard ratio, move with promotion (conservative) | Hazard ratio, move first | Time ratio, switch vs internal |
|---|---|---|---|---|
| SDR to AE or higher | 256,008 people | 1.026 (1.010 to 1.043) | 4.215 | 1.675 (1.660 to 1.689) |
| SDR to AE or higher | education sample, 38,782 | 1.020 (0.979 to 1.063) | 4.114 | 1.676 (1.639 to 1.714) |
| AE to manager or higher | 300,000 people | 1.456 (1.430 to 1.482) | 8.350 | 1.340 (1.322 to 1.358) |
| AE to manager or higher | education sample, 166,952 | 1.434 (1.400 to 1.469) | 8.415 | 1.351 (1.326 to 1.375) |

## 5. Pre-registered hypotheses

1. AE to manager: switch paths 4 to 12 months faster. **Rejected.** Switch paths are 15 months slower.
2. SDR to AE: gap between -2 and +6 months. **Rejected.** The gap is +12.
3. A switch is in more than half of AE-to-manager promotions and less than half of SDR-to-AE promotions. **Supported** (72.5% and 49.7%; the SDR share is just under half).
4. Switch promotions land at smaller companies more often, and normalizing shrinks the switcher advantage by at least a third. **First part supported, second rejected:** there is no switcher advantage to shrink, and the pre-registered rule does not touch the manager step.
5. Adjusted hazard ratio for switch paths above 1 with an interval excluding 1. **Supported for AE to manager** (1.456, conservative coding). For SDR to AE it is 1.026 on the full cohort and its interval includes 1 on the education sample.

## 6. Verification

An independent re-analysis, written from the study plan alone without seeing our code, recomputed the headline numbers. Its first pass matched every median and gap but differed by 0.1 to 0.3 points on route shares and leaving-sales incidence. The causes were three rules in our code: same-month order of sales and non-sales roles, a same-month sales role at another company, and student and intern titles counted as leaving sales. We changed our code to the re-analysis's reading, which follows the plan more closely. After the fix every headline number matches to the precision reported here.

## 7. Limitations

- **Selection, not cause.** People choose to stay or leave. Strong reps get poached; reps passed over for promotion leave. We cannot separate these, so we say "linked with", never "causes".
- **Conditioning on promotion.** The medians compare only people who were promoted. Recent starters have had less time, so only fast promotions show up for them; that is why the gap is smaller for 2020-2023 starters. The cumulative incidence curves use everyone.
- **Self-reported profiles.** People leave short jobs off, add jobs late, and delete profiles. Only people currently in six English-speaking countries are covered. Earlier years undercount.
- **Titles.** Levels come from free-text titles via fixed rules. English titles are classified best. A level is a title, not proof of managing people.
- **Company size** is rebuilt from today's employee count and each company's trend in the extract, not recorded at the time. Size is unknown for 27% of needed company ids with no matching record and for companies known only by name.
- **Company identity.** Where the company id is missing we match by name, so two companies with the same name merge and a renamed company splits.
- **Side roles.** Overlapping roles at other companies (for example advisory roles in a sales function) can mark a path as a switch when the person never left.
- **Month precision.** A move and a promotion in the same month cannot be ordered, which is why the Cox model is fitted two ways.
- **The pre-registered title-inflation rule** does not affect the manager threshold. The checks that do are exploratory.
- **Missing short jobs** were not estimated in this study.

## 8. Sources

- Ken internal data: Ken's dataset of public professional profiles, work histories and company records, accessed 2026-10-16.
- Ken Research shared sales careers extract, built for the GTM engineer study, 2026-10-15.
- Company records pulled by the earlier tenure study, reused.
- Bidwell, M. (2011). Paying more to get less: The effects of external hiring versus internal mobility. Administrative Science Quarterly 56(3), 369-407.
- Bidwell, M. and Mollick, E. (2015). Shifts and ladders: Comparing the role of internal and external mobility in managerial careers. Organization Science 26(6), 1629-1645.