# Methodology: The first 20 hires: engineers first, a seller by hire 7

By Cristian Frunze, Founder of Ken · Oct 28, 2026 · [Ken Research](/research) No. 15 · Founders and startups, Hiring and org design, Sales teams

The full methodology behind [The first 20 hires: engineers first, a seller by hire 7](/research/first-20-hires.md): the data, every method, every number and every limitation. The same text as the [methodology PDF](/research/first-20-hires.pdf).

Data: Ken internal data, Ken's dataset of public professional profiles, work histories and company records.

## 1. Summary

Among 6,951 B2B software startups founded 2012-2022 with headquarters in the US, UK, Canada, Australia, Ireland or New Zealand and at least 20 hires visible in Ken internal data, the median startup made its first engineering hire at hire 2, its first sales hire at hire 7, its first marketing and its first product or design hire at hire 11, its first operations or admin hire at hire 12 and its first customer success or support hire at hire 15. Fewer than half had a people or recruiting hire (25.0%), a finance or legal hire (28.0%) or a data or research hire (41.5%) among their first 20. The median startup took 44 months from its first hire to its 20th. Among startups whose 20th hire came by the end of 2020, those that later reached 100 people took 29 months against 42 for those that stayed under 50 (difference -13, 95% interval -15 to -11), and their first 20 hires held more salespeople (18.4% against 15.3%) and fewer engineers (36.8% against 39.9%). These are links in a comparison, not causes. All numbers describe people who live in these six countries.

## 2. Question and pre-registration

**Question.** For B2B software startups founded 2012-2022, which functions fill hires 1 to 20, in what order, and how long does it take to go from the first hire to the twentieth? How do the startups that kept growing differ from the ones that stalled?

**Pre-registration.** The hypothesis and plan were pre-registered on 2026-10-19, before any data was read. The pre-registered hypotheses:

1. H1. Engineering is the most common function at each of hires 1 to 5, and hires 16-20 hold a smaller engineering share than hires 1-5.
2. H2. The median first-arrival hire numbers run engineering (1-2), sales (3-6), marketing (5-12), customer success and support (8-20), each after the one before; operations and admin falls at 8-20; fewer than half of the startups have a people or recruiting hire, or a finance or legal hire, in their first 20; the most common order in which the four groups first appear is engineering, sales, marketing, customer success and support.
3. H3. The median time from hire 1 to hire 20 is 24-48 months, and the median company age at hire 20 is 36-60 months.
4. H4. Startups that reached 100 people, against those that stayed under 50: (a) went from hire 1 to hire 20 in less than half the stallers' median time; (b) had an engineering share of hires 1-20 at least 5 points higher and a lower sales share; (c) more often had a people or recruiting hire among the first 20.
5. H5. Removing a quarter of the hires at random does not change the order of the median first-arrival hire numbers of the five core groups, and moves no group's share of hires 1-20 by 3 points or more.

## 3. Data

| Source | What was used | Size |
|---|---|---|
| Ken internal data, people | Every work-history entry (title, company id, start and end month, current flag) of every profile, read from Ken's shared person-level extract (built 2026-10-16/17). The data holds people whose current country is the US, UK, Canada, Australia, Ireland or New Zealand; their earlier jobs can be anywhere. Profiles were last refreshed between 2025-11 and 2026-04. | 246,725,412 profiles in the index; 555,053,752 entries read, of which 2,423,875 at frame companies |
| Ken internal data, company records | Industry label, founded year, HQ country, record type and employee count | 2,726,653 records with HQ in the six countries founded 2012-2022 |
| SEC Form D (public filings), through Ken's shared funding-events extract | First Form D raise per US company, for one robustness row | 741 sample companies with a raise by their 20th hire |
| Ken Research study *When should you hire your first salesperson?* | Published outputs, cited unchanged (Table 8) | - |

## 4. Sample

- **Companies:** records with a product-software industry label (software development; computer and network security; data infrastructure and analytics; business intelligence platforms; data security software products; embedded software products; mobile computing software products; desktop computing software products), founded 2012-2022, HQ in the six countries (88,034 records). Records typed as public company, government agency, nonprofit, educational institution, self-employed or sole proprietorship were dropped, which leaves 78,981. Company records carry no B2B flag; product-software labels are the closest available match, and they include some consumer software.
- **Analysis sample:** companies with 20 or more hires (Section 5) and fewer than 3 people who joined before the founding year: 6,951 companies (5,002 with HQ in the US), holding 609,278 hires.
- **Outcome set:** sample companies whose 20th hire joined by 2020-12, so that each had more than five years to grow by 2026 Q1: 2,720 companies. **Grew:** rebuilt headcount reached 100 at some quarter end by 2026 Q1 (606). **Stalled:** rebuilt headcount stayed under 50 at every quarter end through 2026 Q1 (1,462). In between (652) are reported but not compared.

## 5. Definitions

**People and join month.** Everyone with a work-history entry at the company (matched on company id). Intern, programme and non-staff entries are set aside first; a person's join month is the start month of their earliest remaining dated entry there, and their function is that entry's title.

**Founders** are tagged and never counted as hires: anyone with a founder title on any entry at the company (founder, co-founder, founding partner, owner; "product owner" and similar job titles excluded), and anyone whose earliest dated entry there (of any kind, intern and advisor entries included) starts before the end of the founding year with a chief executive, other C-level or managing director title on an entry that starts that month. 12,289 founders were tagged at the sample companies (10,084 by title, 2,205 by a C-level title in the founding year); 1,004 companies have no tagged founder, usually because the founders are not in the data.

**Not hires.** Interns and programme roles (fellowships, students, MBA and summer roles), non-staff titles (advisors, board members, investors, mentors, ambassadors, volunteers, freelancers and gig listings), people with no dated entry, people who joined before 1 January of the founding year, and people who joined after 2026-03 (the last complete quarter, since profiles add new jobs late).

**Hire order.** Hires are ordered by join month. Dates are month precision, and 46.3% of hires 1-20 share their join month with another hire at the same company; ties are broken by a fixed pseudo-random key of the person id, (id x 2654435761) mod 2^32, then by the person id, so tied hires spread evenly over their ranks.

**Function groups.** Titles are labelled with the shared Ken Research title taxonomy (ordered regular expressions on the normalized title) and grouped into 11 functions: engineering; product and design; data and research; sales (including business development, account management, sales engineering and revenue operations); marketing; customer success and support; operations and admin (including chiefs of staff, project and program management and strategy roles); people and recruiting; finance and legal; general management (hired chief executives, presidents, general and managing directors); and unclassified titles. A study layer relabels titles the shared taxonomy leaves unclassified. It was fixed from an audit of the 600 most common unclassified titles among hires 1-20, before any comparison of outcomes; it cut the unclassified share of hires 1-20 from 9.16% to 5.23%.

**Rebuilt headcount.** A person counts at a company at a quarter end when one of their entries there started on or before it and ended on or after it (the end month counts in full) or is current. Everyone counts, founders included. Quarters run from 2010 Q1 to 2026 Q1. From 2015 Q1 all 2,257,275 company-quarters equal Ken's shared quarterly headcount table.

**Company age** is counted from 1 July of the founding year, because records give only the year.

## 6. Methods

1. **Function of hire n:** the share of companies whose hire n (n = 1-20) is in each group, and the same pooled over hires 1-5, 6-10, 11-15, 16-20 and 1-20.
2. **First arrivals:** for each group and n, the share of companies with at least one hire in the group among hires 1 to n. The median first-arrival hire number is the smallest n at which that share reaches 50%, which equals the ceil(N/2)-th smallest first-arrival hire number when companies with none among their first 20 count as later than 20; "after 20" when the share never reaches half.
3. **Pace:** months from founding to hires 1, 10 and 20, and from hire 1 to hire 20; medians and quartiles, overall and by founding cohort.
4. **Sequences:** the order in which engineering, sales, marketing and customer success and support (and, separately, those four plus operations and admin) first appear among hires 1-20, and the groups of hires 1-3.
5. **Grew against stalled:** methods 1-3 for each outcome group, and the differences with 95% percentile intervals from 1,000 bootstrap resamples of companies within each group (seed 71). A robustness version reweights the stalled group to the growers' mix of founding year and HQ (US against the other five countries).
6. **Missing and deleted profiles:** a count of profiles flagged deleted in the people index; random removal of 10%, 25% and 50% of each company's hires (50 draws each); removal of 25% of all non-engineering hires, and of 25% of engineering hires; dropping jobs that ended within 6 months; founding cohorts; companies whose rebuilt headcount is 0.7-1.3 times the employee count on their record.
7. **Other robustness:** founders by title only; tied hires ranked together; the "technology, information and internet" label added; public companies kept; US companies with a Form D raise by their 20th hire; US against the other five countries.
8. **Privacy:** only aggregate tables leave Ken's servers; every published count under 50 companies or people is shown as "<50", and statistics resting on fewer than 50 are suppressed.

## 7. Results

### 7.1 Who fills the first 20 seats

Engineering is the most common group at each of hires 1 to 5. Its share falls from 41.1% of hires 1-5 to 34.0% of hires 16-20, while sales, marketing and customer success together rise from 27.1% to 37.2% (Table 2, Figure 2).

### 7.2 First arrivals

Half of the startups have an engineer by hire 2, a salesperson by hire 7, a marketer and a product or design hire by hire 11, an operations or admin hire by hire 12 and a customer success or support hire by hire 15 (Figure 1, Table 2). Data and research, people and recruiting, finance and legal, and general management never reach half within 20 hires.

### 7.3 Sequences

Among the 2,655 startups whose first 20 hires include engineering, sales, marketing and customer success and support, the most common first-appearance order is Eng > Sales > Mkt > CS (11.5%), then Eng > Sales > CS > Mkt (11.2%) and Eng > Mkt > Sales > CS (10.7%) (Table 7). 12.4% of all sample startups made three engineering hires first.

### 7.4 Pace

From founding, hire 10 comes after a median 31 months (middle half 19-47) and hire 20 after 49 (34-71). Hire 1 to hire 20 takes 44 months (30-62): 56 for companies founded 2012-2014, 47 for 2015-2017, 37 for 2018-2020 and 30 for 2021-2022 (Table 3). Younger cohorts only hold companies that reached 20 hires quickly, and older cohorts have probably lost more early profiles, so the trend across cohorts is not a clean measure of change.

### 7.5 Grew against stalled

Growers took 29 months from hire 1 to hire 20 against 42 (-13, -15 to -11; ratio 0.69, 0.65 to 0.73). Their first 20 hires held 18.4% sales against 15.3% (+3.1 points, +1.7 to +4.6) and 36.8% engineering against 39.9% (-3.1, -4.8 to -1.3). 29.9% of growers had a people or recruiting hire among their first 20, against 23.1% (+6.8, +2.8 to +11.3), and 34.0% a finance or legal hire, against 26.5%. Reweighting the stallers to the growers' founding years and US share gives a pace gap of -10 months (-12 to -8) and a sales gap of +3.3 points (+1.9 to +4.8) (Table 4, Figure 3). The comparison shows a link, not a cause.

### 7.6 The first sales hire, cited from the first sales hire study

The first sales hire study measured the first account executive at a median 47 months after founding and 12 people on the team; 36 months and 28 people at companies that passed 100, 49 months and 11 people at companies that stayed under 100 (Table 8). Its sample differs (software and IT services labels, founded 2010-2020, a rebuilt headcount of 10 or more), and its account executive is a narrower role than this study's sales group, so the two measures are not the same.

## 8. Pre-registered hypotheses: verdicts

| Hypothesis | Result | Verdict |
|---|---|---|
| H1 Engineering first, go-to-market later | Engineering most common at each of hires 1-5; engineering share 41.1% (hires 1-5) to 34.0% (16-20) | holds |
| H2 Order and ranges | Engineering 2, sales 7, marketing 11, customer success and support 15, operations and admin 12; people 25.0% and finance and legal 28.0% in the first 20; most common order Eng > Sales > Mkt > CS (11.5%) | wrong: sales at hire 7 is outside 3-6; every other part holds |
| H3 Pace | 44 months from hire 1 to hire 20; 49 months old at hire 20 | holds |
| H4a Growers under half the time | ratio 0.69 (0.65 to 0.73) | wrong |
| H4b Growers more engineering, less sales | engineering -3.1 points, sales +3.1 points | wrong, in the opposite direction |
| H4c Growers more often a people hire | +6.8 points (+2.8 to +11.3) | holds |
| H5 Order robust to 25% removal | no order change; largest share move 1.0 points | holds |

## 9. Missing and deleted profiles

The brief asks how much missing or deleted early profiles could shift the hire order. Ken internal data holds only profiles that exist today: the people index has 246,725,412 profiles and 0 flagged as deleted, because deleted profiles leave the index entirely. Their number cannot be counted, so the analysis removes hires on purpose and watches what moves (Table 5).

- **Order.** No removal changes the order of first arrivals of the five core groups. At 25% random removal the first sales hire moves from 7 to 6 and customer success and support from 15 to 14; marketing (11) and operations (12) do not move. No group's share of hires 1-20 moves by more than 1.0 points.
- **Hire numbers compress.** Each missing early hire pulls later hires forward, so measured hire numbers run early. Run in reverse, the same logic says that if a quarter of early hires are missing, the true hire numbers sit about one position later than measured for sales and customer success. Removing a quarter of every function except engineering, the case where engineers are the best covered, pushes marketing to 13, operations to 14 and customer success to 16.
- **Pace stretches.** Fewer visible hires means the 20th visible hire comes later: on the same companies, 25% removal stretches hire 1 to hire 20 from 40 to 46 months and 50% removal from 36 to 51. The measured 44 months is therefore likely an upper end of the true time to 20 hires.
- **Short jobs.** Dropping the 43,439 hires whose job ended within 6 months moves sales to hire 6 and customer success to 14.

## 10. Robustness

Across the variants in Table 6 (founding cohorts, coverage, the founder rule, the tie rule, the industry set, public companies, US and non-US headquarters, US companies with a Form D raise (741) and the three outcome groups), the median first-arrival hire number ranges over 2 for engineering, 6-7 for sales, 10-13 for marketing, 9-12 for product and design, 11-13 for operations and admin and 13-17 for customer success and support. Companies whose rebuilt headcount is close to the employee count on their record (across the sample the median ratio is 0.62) keep the order of engineering, sales, marketing and customer success, with the first marketer later (hire 13 against 10 for the rest).

## 11. Verification

An independent re-analysis, written without seeing our code, recomputed 88 of the study's numbers from the same extracts. It was given the question, the brief, the hypotheses, the plan and decisions D1-D25, plus the shared title definitions, and no study code or outputs. It rebuilt the founders, the hires, the hire order and the quarterly headcount itself and reran the grew and stalled bootstrap and the 25% removal simulation. The match rules were fixed before its numbers were read (D25: counts and medians equal, shares to 0.1 percentage point, interval ends within a quarter of the interval width, simulation means within 0.3 hires, 0.3 points or 1 month). In the first comparison 82 of 88 numbers matched. Three counts differed by a few people: founders 12,226 against 12,289 in the re-analysis, hires 609,280 against 609,278, and companies with all four core functions among their first 20 hires 2,656 against 2,655. Three shares differed in the third decimal only because the 4-decimal value from the re-analysis, rounded again, crossed a half-unit boundary; each agrees with the study at 4 decimals, which is now also accepted as a match (D27). The count differences came from the founder rule. The plan calls a person a founder when their first spell at the company is a C-level role starting in the founding year or earlier; the study had judged the first staff spell, after intern and advisor spells were set aside, while the re-analysis judged the earliest spell of any kind. The plan's wording supports the reading of the re-analysis, so the study adopted it, added the person id as the last tie-breaker as the re-analysis had, and reran (D26). In the second comparison all 88 numbers matched; three shares among them match at 4 decimals rather than 3. The fix raised the founder count by 63, net, and left the sample and the outcome groups unchanged; no share in the result tables moved by more than 0.2 percentage points, one median moved by one hire (the stalled companies' first marketing hire, 11 to 12), and no verdict changed. The re-analysis did not repeat the check of the rebuilt headcount against the shared headcount table, which was not among its inputs; the study's own check found the two equal in every cell. It also reran readings the plan leaves open: the two other founder readings give 12,231 and 12,225 founders instead of 12,289 and change no median; taking the other title when a person has two entries starting in their join month moves no share by more than 0.2 points; counting pre-founding joiners over spells of any kind gives a sample of 6,842 companies instead of 6,951 and changes no median; leaving intern and non-staff spells out of the rebuilt headcount gives 564 grew and 1,538 stalled companies instead of 606 and 1,462, and a pace gap of 12 months instead of 13.

## 12. Alternative explanations considered

1. **Missing profiles, not hiring order.** Early staff who deleted their profile or left a short job off it are missing. Simulated removal (Section 9) changes no order and moves hire numbers by 1-3; the order is robust, the hire numbers and the pace are approximate.
2. **Staff outside the six countries.** Offshore engineers or support staff are missing, so engineering and support may arrive earlier than measured at startups with teams abroad. Companies whose visible headcount is close to their recorded size give the same order of engineering, sales, marketing and customer success.
3. **Founders counted as hires, or hires counted as founders.** 1,004 companies have no tagged founder; their first visible people may be founders with plain titles. Founders by title only (7g) gives the same medians.
4. **Title classification.** 5.2% of hires 1-20 stay unclassified, and borderline titles (business analysts, project managers, consultants) could sit in another group. The order of the main groups does not depend on these.
5. **Same-month starts.** 46.3% of early hires share a month; ranking ties together (7h) moves sales to 6, operations to 11 and customer success to 14.
6. **Cohort and survivorship.** Growers and stallers differ in founding year and country. Reweighting keeps the direction of the pace, sales, engineering and people differences; a startup's demand may drive both its hiring and its growth, so the growth comparison is a link, not a cause. Acquired companies stop growing in the data and can look like stallers.
7. **Industry labels.** Product-software labels include some consumer software; adding the broad internet label changes little.

## 13. Limitations

- Ken internal data covers people who currently live in the US, UK, Canada, Australia, Ireland and New Zealand. Staff elsewhere, and people without a public profile, are missing.
- Profiles are self-reported. Short jobs are often left off, titles are sometimes edited after a promotion, and deleted profiles are gone.
- Profiles add new jobs months late; joins after 2026-03 are not counted, and the latest quarters undercount.
- Dates have month precision; the order within a month is unknown.
- Many work-history entries in Ken internal data carry no company id and cannot be placed at a company; those people are missing from every company's team.
- Company records give only the founding year; age is measured from 1 July.
- "Grew" and "stalled" use visible headcount in the six countries, not total employees.
- The outcome comparison is observational.

## 14. Privacy

Only aggregate tables leave Ken's servers. No person or company is named. Counts under 50 are shown as "<50", statistics resting on fewer than 50 companies are suppressed, and sequences held by fewer than 50 companies are pooled.

## 15. Sources and reuse

- Ken internal data (people and company records), through Ken's shared headcount history (built for the Ken Research study *Did AI kill the SDR role?*) and the shared title taxonomy (built for the Ken Research study *Who are GTM engineers, and where do they come from?*).
- SEC Form D data sets, through Ken's shared funding-events extract (built for the first sales hire study).
- The first sales hire study: its first-hire timing results, cited unchanged in the last table below.
- The Ken Research study *Sales and marketing draw level with engineering at about 50 people*: product-software list, sales group, non-staff words, tie key, well-covered band.
- Decisions D1-D27, each logged with its reason in the study plan.

## Tables

### Table 1. Sample flow

| Step | Companies |
|---|---|
| product-software records, six-country HQ, founded 2012-2022 | 88,034 |
| after the record-type filter (D13) | 78,981 |
| with at least one person in Ken internal data | 52,086 |
| with at least one hire | 39,850 |
| dropped: 3 or more joiners before the founding year (D9) | 1,844 |
| analysis sample: 20 or more hires | 6,951 |
| outcome set: 20th hire by 2020-12 (D8) | 2,720 |
| grew: rebuilt headcount reached 100 by 2026 Q1 | 606 |
| stalled: rebuilt headcount under 50 through 2026 Q1 | 1,462 |
| in between: peak 50-99 | 652 |

### Table 2. Function of hires 1-20 (6,951 companies)

| Function group | Share of hires 1-5 | Hires 16-20 | Hires 1-20 | Companies with one in hires 1-20 | Median first-arrival hire number | Middle half, if one in hires 1-20 |
|---|---|---|---|---|---|---|
| Engineering | 41.1% | 34.0% | 37.0% | 96.6% | 2 | 1-3 |
| Product and design | 8.5% | 7.4% | 7.9% | 67.4% | 11 | 3-11 |
| Data and research | 5.2% | 5.2% | 5.2% | 41.5% | after 20 | 3-12 |
| Sales | 13.7% | 20.2% | 17.4% | 82.4% | 7 | 2-9 |
| Marketing | 7.6% | 8.1% | 8.0% | 68.6% | 11 | 3-12 |
| Customer success and support | 5.8% | 8.9% | 7.8% | 59.5% | 15 | 4-12 |
| Operations and admin | 7.1% | 6.6% | 6.7% | 67.4% | 12 | 3-12 |
| People and recruiting | 1.5% | 2.3% | 1.9% | 25.0% | after 20 | 5-15 |
| Finance and legal | 2.0% | 2.2% | 2.0% | 28.0% | after 20 | 4-15 |
| General management | 1.2% | 0.6% | 0.8% | 12.6% | after 20 | 3-13 |
| Unclassified title | 6.3% | 4.6% | 5.2% | 51.4% | 19 | 2-12 |

### Table 3. Pace, months: median (middle half)

| Group | Companies | Founding to hire 10 | Founding to hire 20 | Hire 1 to hire 20 |
|---|---|---|---|---|
| All | 6,951 | 31 (19-47) | 49 (34-71) | 44 (30-62) |
| All, founded 2012-2014 | 2,053 | 42 (25-63) | 66 (42-95) | 56 (37-81) |
| All, founded 2015-2017 | 2,439 | 34 (22-49) | 54 (38-72) | 47 (34-64) |
| All, founded 2018-2020 | 1,860 | 25 (16-35) | 41 (28-56) | 37 (26-50) |
| All, founded 2021-2022 | 599 | 18 (10-26) | 33 (22-42) | 30 (21-39) |
| Grew to 100 | 606 | 21 (13-31) | 32 (22-43) | 29 (21-39) |
| Peaked at 50-99 | 652 | 25 (16-35) | 38 (27-50) | 35 (24-45) |
| Stayed under 50 | 1,462 | 28 (18-41) | 47 (33-62) | 42 (30-56) |

### Table 4. Grew (606) against stalled (1,462)

| Statistic | Grew | Stalled | Difference (95% interval) | Reweighted difference |
|---|---|---|---|---|
| Months, hire 1 to hire 20 | 29 | 42 | -13 (-15 to -11) | -10 (-12 to -8) |
| Months, founding to hire 10 | 21 | 28 | -7 (-9 to -5) | -4 (-5 to -2) |
| Months, founding to hire 20 | 32 | 47 | -15 (-17 to -13) | -11 (-13 to -9) |
| Share of hires 1-20: Engineering | 36.8% | 39.9% | -3.1 (-4.8 to -1.3) pts | -2.7 (-4.5 to -0.7) pts |
| Share of hires 1-20: Product and design | 7.3% | 8.7% | -1.4 (-2.2 to -0.6) pts | -1.7 (-2.6 to -0.9) pts |
| Share of hires 1-20: Data and research | 4.9% | 5.4% | -0.5 (-1.4 to +0.3) pts | -0.7 (-1.7 to +0.2) pts |
| Share of hires 1-20: Sales | 18.4% | 15.3% | +3.1 (+1.7 to +4.6) pts | +3.3 (+1.9 to +4.8) pts |
| Share of hires 1-20: Marketing | 7.6% | 7.9% | -0.3 (-1.2 to +0.5) pts | -0.3 (-1.2 to +0.6) pts |
| Share of hires 1-20: Customer success and support | 7.9% | 6.7% | +1.1 (+0.2 to +2.0) pts | +1.3 (+0.3 to +2.2) pts |
| Share of hires 1-20: Operations and admin | 6.0% | 6.8% | -0.8 (-1.4 to -0.1) pts | -0.8 (-1.5 to -0.1) pts |
| Share of hires 1-20: People and recruiting | 2.2% | 1.8% | +0.4 (0.0 to +0.8) pts | +0.1 (-0.5 to +0.6) pts |
| Share of hires 1-20: Finance and legal | 2.7% | 1.7% | +1.0 (+0.6 to +1.6) pts | +1.1 (+0.6 to +1.6) pts |
| Share of hires 1-20: General management | 0.6% | 0.8% | -0.2 (-0.4 to 0.0) pts | -0.2 (-0.3 to 0.0) pts |
| Share of hires 1-20: Unclassified title | 5.6% | 5.0% | +0.6 (-0.1 to +1.5) pts | +0.6 (-0.2 to +1.5) pts |
| Companies with one in hires 1-20: Engineering | 97.7% | 97.5% | +0.2 (-1.2 to +1.7) pts | +0.6 (-1.0 to +2.2) pts |
| Companies with one in hires 1-20: Product and design | 67.5% | 70.0% | -2.5 (-6.9 to +2.0) pts | -3.3 (-7.9 to +1.3) pts |
| Companies with one in hires 1-20: Data and research | 40.4% | 41.6% | -1.2 (-5.9 to +3.7) pts | -1.6 (-6.1 to +3.6) pts |
| Companies with one in hires 1-20: Sales | 82.0% | 79.8% | +2.2 (-1.1 to +5.9) pts | +2.7 (-1.1 to +6.5) pts |
| Companies with one in hires 1-20: Marketing | 67.8% | 68.0% | -0.2 (-5.0 to +4.3) pts | -0.1 (-5.1 to +4.4) pts |
| Companies with one in hires 1-20: Customer success and support | 62.5% | 56.5% | +6.0 (+1.4 to +10.7) pts | +7.0 (+2.0 to +11.6) pts |
| Companies with one in hires 1-20: Operations and admin | 64.9% | 67.4% | -2.6 (-7.1 to +2.0) pts | -2.9 (-7.6 to +2.0) pts |
| Companies with one in hires 1-20: People and recruiting | 29.9% | 23.1% | +6.8 (+2.8 to +11.3) pts | +5.4 (+1.1 to +10.1) pts |
| Companies with one in hires 1-20: Finance and legal | 34.0% | 26.5% | +7.5 (+3.3 to +12.0) pts | +8.1 (+3.8 to +12.6) pts |
| Companies with one in hires 1-20: General management | 10.2% | 13.5% | -3.3 (-6.1 to 0.0) pts | -2.8 (-5.6 to +0.5) pts |
| Companies with one in hires 1-20: Unclassified title | 54.3% | 50.6% | +3.7 (-0.9 to +8.8) pts | +3.2 (-1.6 to +8.5) pts |
| Median first-arrival hire number: Engineering | 2 | 2 | 0 (0 to 0) | 0 (0 to 0) |
| Median first-arrival hire number: Product and design | 11 | 10 | 1 (-2 to 2) | 1 (-1 to 3) |
| Median first-arrival hire number: Sales | 6 | 7 | -1 (-2 to 0) | -1 (-2 to 0) |
| Median first-arrival hire number: Marketing | 12 | 12 | 0 (-1 to 2) | 0 (-2 to 2) |
| Median first-arrival hire number: Customer success and support | 13 | 17 | -4 (-6 to 0) | -4 (-7 to -1) |
| Median first-arrival hire number: Operations and admin | 13 | 12 | 1 (-1 to 3) | 1 (-1 to 4) |
| Median first-arrival hire number: Unclassified title | 18 | 20 | -2 (-6 to 1) | -2 (-6 to 2) |

### Table 5. Missing profiles: median first-arrival hire number and months from hire 1 to hire 20 after removing hires (means over 50 draws)

| Hires removed | Companies left (mean) | Sales | Marketing | Operations and admin | Customer success and support | Hire 1 to 20, months | Same companies, full data |
|---|---|---|---|---|---|---|---|
| none (full data) | 6,951 | 7 | 11 | 12 | 15 | 44 | 44 |
| 10% at random | 6,476 | 6.8 | 11.0 | 12.0 | 14.3 | 45.0 | 42.0 |
| 25% at random | 5,420 | 6.0 | 11.0 | 12.0 | 14.0 | 46.0 | 40.0 |
| 50% at random | 3,865 | 5.9 | 11.0 | 12.7 | 12.1 | 51.1 | 36.0 |
| 25% of every function except engineering | 6,048 | 7.0 | 13.0 | 13.9 | 16.0 | 45.4 | 41.0 |
| 25% of engineering only | 6,435 | 6.0 | 10.0 | 10.5 | 13.0 | 44.8 | 42.0 |
| jobs that ended within 6 months | 6,402 | 6.0 | 11.0 | 12.0 | 14.0 | 44.0 | 42.0 |

### Table 6. Robustness: median first-arrival hire number by variant

| Variant | Companies | Engineering | Sales | Marketing | Product and design | Operations and admin | Customer success and support | Hire 1 to 20, months |
|---|---|---|---|---|---|---|---|---|
| main | 6,951 | 2 | 7 | 11 | 11 | 12 | 15 | 44 |
| 6c short jobs dropped | 6,402 | 2 | 6 | 11 | 11 | 12 | 14 | 44 |
| 6d founded 2012-2016 | 3,665 | 2 | 6 | 11 | 12 | 12 | 13 | 52 |
| 6d founded 2017-2022 | 3,286 | 2 | 7 | 12 | 9 | 12 | 16 | 37 |
| 6e well covered (0.7-1.3) | 2,673 | 2 | 7 | 13 | 12 | 12 | 15 | 45 |
| 6e rest | 4,278 | 2 | 7 | 10 | 10 | 12 | 15 | 43 |
| 7g founders by title only | 7,027 | 2 | 7 | 11 | 11 | 12 | 15 | 44 |
| 7h ties ranked together | 6,951 | 2 | 6 | 11 | 10 | 11 | 14 | 44 |
| 7i technology, information and internet added | 9,139 | 2 | 7 | 10 | 10 | 11 | 15 | 43 |
| 7k public companies kept | 7,165 | 2 | 7 | 11 | 11 | 12 | 15 | 44 |
| 7j US, Form D raise by hire 20 | 741 | 2 | 6 | 10 | 10 | 12 | 14 | 40 |
| HQ US | 5,002 | 2 | 6 | 11 | 11 | 12 | 15 | 42 |
| HQ other five countries | 1,949 | 2 | 7 | 11 | 11 | 12 | 15 | 48 |
| outcome grew | 606 | 2 | 6 | 12 | 11 | 13 | 13 | 29 |
| outcome stalled | 1,462 | 2 | 7 | 12 | 10 | 12 | 17 | 42 |
| outcome middle | 652 | 2 | 7 | 11 | 11 | 11 | 14 | 35 |

### Table 7. Order in which engineering, sales, marketing and customer success first appear, companies with all four in hires 1-20 (2,655)

| Order | Companies | Share |
|---|---|---|
| Eng > Sales > Mkt > CS | 304 | 11.5% |
| Eng > Sales > CS > Mkt | 297 | 11.2% |
| Eng > Mkt > Sales > CS | 285 | 10.7% |
| Eng > CS > Sales > Mkt | 220 | 8.3% |
| Eng > Mkt > CS > Sales | 183 | 6.9% |
| Sales > Eng > Mkt > CS | 162 | 6.1% |
| Sales > Eng > CS > Mkt | 152 | 5.7% |
| Eng > CS > Mkt > Sales | 140 | 5.3% |
| Mkt > Eng > Sales > CS | 112 | 4.2% |
| CS > Eng > Sales > Mkt | 84 | 3.2% |
| Sales > CS > Eng > Mkt | 82 | 3.1% |
| Sales > Mkt > Eng > CS | 78 | 2.9% |
| Mkt > Eng > CS > Sales | 71 | 2.7% |
| Mkt > Sales > Eng > CS | 66 | 2.5% |
| Sales > Mkt > CS > Eng | 64 | 2.4% |
| Sales > CS > Mkt > Eng | 56 | 2.1% |
| Mkt > Sales > CS > Eng | 50 | 1.9% |
| other orders (7 orders, each under 50 companies) | 249 | 9.4% |

### Table 8. Cited from the first sales hire study

| First sales hire study group | First hire | Metric | Median | Companies |
|---|---|---|---|---|
| all | First AE | age_months | 47 | 9,970 |
| all | First AE | headcount_before_hire_quarter | 12 | 9,970 |
| all | First marketing hire | age_months | 32 | 18,995 |
| all | First marketing hire | headcount_before_hire_quarter | 8 | 18,995 |
| all | First sales leader | age_months | 44 | 12,611 |
| all | First sales leader | headcount_before_hire_quarter | 11 | 12,611 |
| all | First SDR | age_months | 54 | 7,459 |
| all | First SDR | headcount_before_hire_quarter | 16 | 7,459 |
| passed_100 | First AE | age_months | 36 | 1,601 |
| passed_100 | First AE | headcount_before_hire_quarter | 28 | 1,601 |
| stalled | First AE | age_months | 49 | 8,369 |
| stalled | First AE | headcount_before_hire_quarter | 11 | 8,369 |