# Methodology: Where does Series A money go? As much to selling as to building

By Cristian Frunze, Founder of Ken · Oct 18, 2026 · [Ken Research](/research) No. 10 · Founders and startups, Hiring and org design, Sales teams

The full methodology behind [Where does Series A money go? As much to selling as to building](/research/series-a-hiring.md): the data, every method, every number and every limitation. The same text as the [methodology PDF](/research/series-a-hiring.pdf).

Data: Ken internal data (Ken's dataset of public professional profiles, work histories and company records) and the SEC's public Form D data sets. Analysis run on 2026-10-19. The hypothesis and plan were pre-registered on 2026-10-19, before any data was pulled. Counts of 1 to 49 companies or people are shown as "<50".

## 1. Summary

**Question.** In the 12 months after a Series A or Series B, which functions do B2B companies hire into? How does that mix differ for companies that went on to raise again?

**Answer.** In the 12 months after a Series A, the 699 US-headquartered B2B software companies in the sample added 13.4 people each on average (95% interval 11.8 to 15.2), against 7.5 in the 12 months before. Of those net adds, 27.0% went to building the product (engineering 21.5%, product 3.7%, data 1.8%) and 26.9% to selling it (sales 19.8%, marketing 7.1%); customer success and support took 7.7%, recruiting and HR 2.0%, operations 2.6% and titles the function rules do not classify 33.8%. The mix in the year before the round was similar: the sales and marketing share changed by +2.8 points (-0.4 to 5.8). After a Series B (272 companies) the sales and marketing share was 30.2% (difference from a Series A +3.3 points, 0.2 to 7.6); within size bands the gap is about as large but not clear (+3.1 points at 5-24 people, +2.2 at 25-99), so size does not explain it away. 50.6% of Series A companies followed for 36 months filed another Form D raise; they had added 17.5 people in the year after the A against 10.5 (+7.1, 3.6 to 11.2), with the same sales and marketing share (26.6% against 27.1%; standardized difference +0.8, -3.6 to 5.7). All counts are of staff in six countries, and these are correlations around dated events, not the effect of the money.

## 2. Data

- **Ken internal data, people.** Public professional profiles of people whose current country is the US, UK, Canada, Australia, Ireland or New Zealand, with every role on the profile: title, company id, start and end month, and a current flag. Profiles were last refreshed between 2025-11 and 2026-04. Earlier roles can be anywhere, but people who now live outside the six countries are not in the data.
- **Ken internal data, companies.** Company records with industry label, founded year, headquarters country and the employee count on the record (a worldwide count).
- **Shared headcount history** (built for the Ken Research study *Did AI kill the SDR role?*, build of 2026-10-17). It counts, at every quarter end from 2015 Q1 to 2026 Q1, the people whose role at a company covers that date: started on or before it, and ended on or after it (end month counted in full) or still current (current roles run to the 2026-04-30 snapshot). Roles without a start date are dropped. Every title carries a function and a level from the shared sales title taxonomy (built for the Ken Research study *Who are GTM engineers, and where do they come from?*) plus a customer support rule. This study read its company records, its company-quarter tables by function and level, and, for section 7.7, its person-level role table.
- **Shared funding events extract** (built for the Ken Research study *When should you hire your first salesperson?*, build of 2026-10-16). Every SEC Form D offering in the SEC's quarterly data sets from 2008 Q1 to 2026 Q3 by a US operating company (pooled funds, SPV-style issuers and non-US issuers removed), one row per offering (issuer and SEC file number, so amendments fold into the offering), dated by its first filing. Offerings are matched to company records on normalized legal name and headquarters state, or on name and incorporation year when the record has no state; names shared by two records in a state stay unmatched. It also carries the last round printed on each company record (round type and date).
- No new data was pulled. Both shared extracts were used read-only.

## 3. Sample

**Frame.** Company records with a US headquarters, founded 2005-2024, in eight product-software industry labels (the set used in the Ken Research study *Sales and marketing draw level with engineering at about 50 people*). 82,800 records:

| Industry label | Records |
|---|---|
| software development | 70,480 |
| computer and network security | 7,345 |
| data infrastructure and analytics | 2,174 |
| business intelligence platforms | 742 |
| data security software products | 436 |
| embedded software products | 423 |
| mobile computing software products | 776 |
| desktop computing software products | 424 |

For robustness (d), 145,826 records labelled "technology, information and internet", "IT services and IT consulting" or "IT system custom software development" were pulled too.

**Form D coverage.** 3,289 of the 82,800 B2B software records (4.0%) have at least one matched Form D offering, 2,638 with one dated 2015-2024. Of all 147,604 Form D offerings by US operating companies dated 2015-2024, 33,810 (22.9%) are matched to a company record; in the Form D technology industry groups (Computers, Other Technology, Telecommunications) 15,035 of 38,723 (38.8%). By issuer, 15.4% of all and 33.8% of technology issuers are matched. Not every round is filed, and the filings cover US issuers only.

**Rounds.** The 13,440 Form D offerings of frame companies (both industry sets) were merged into 12,569 rounds at 5,850 companies: an offering joins a round when it was first filed within 90 days of the round's first filing. Offering amount = amount sold if above zero, else the offering amount. A round is a share sale (equity) when any of its offerings is marked as equity.

**Events.** One event per company and series. Sample flow under rule R1:

| Series | Step | Events |
|---|---|---|
| Series A | labelled round dated 2015-04-01 to 2024-12-31 (frame, both industry sets) | 1512 |
| Series A | dated 2016-04-01 or later | 1375 |
| Series A | product-software industry | 942 |
| Series A | company has rebuilt headcount in some quarter | 831 |
| Series A | headcount at base quarter 5 or more | 699 |
| Series A | headcount at base quarter 500 or less (main sample) | 699 |
| Series A | main sample, dated up to 2023-09-30 (raised-again comparison) | 628 |
| Series A | robustness (i): dated 2015-04-01 to 2016-03-31 | 63 |
| Series A | robustness (d): broader industry set added | 933 |
| Series B | labelled round dated 2015-04-01 to 2024-12-31 (frame, both industry sets) | 470 |
| Series B | dated 2016-04-01 or later | 423 |
| Series B | product-software industry | 306 |
| Series B | company has rebuilt headcount in some quarter | 280 |
| Series B | headcount at base quarter 5 or more | 273 |
| Series B | headcount at base quarter 500 or less (main sample) | 272 |
| Series B | main sample, dated up to 2023-09-30 (raised-again comparison) | 249 |
| Series B | robustness (i): dated 2015-04-01 to 2016-03-31 | <50 |
| Series B | robustness (d): broader industry set added | 335 |

Median Series A amount $8.0M (middle half $5.4M to $12.7M); median Series B $22.9M ($15.4M to $37.3M). Median rebuilt headcount at the base quarter: 14 people (Series A) and 38 (Series B).

## 4. Classifying Series A and Series B

Form D does not name the series, so rounds are labelled by amount and order.

- **Rule R1 (headline).** Series A = the company's first round of $4M or more, if that round is a share sale of at most $30M (otherwise the company has no Series A under R1). Series B = the first later share sale of at least $10M and at least 1.5 times the Series A amount, and at most $100M. Rounds before the A and between the A and the B are seed rounds, extensions or bridges. The thresholds follow Carta's published round sizes for 2021-2024 (median Series A about $7-13M, median Series B about $12-23M, seed rounds up to about $5M at the 75th percentile).
- **Rule R2 (robustness).** Series A = the first share sale of $4M to under $20M with no earlier round of $4M or more; Series B = the first share sale of $20M to under $80M with no earlier round of $20M or more.

**Check against company records.** Each frame company whose record has a last round dated 2015-2024 was paired with its Form D round nearest the record date within 180 days. Record types were grouped into seed (Pre seed, Seed, Angel), A, B, C+ (Series C to J) and other. Precision = the share of rule-labelled rounds, among those paired with a record of type seed, A, B or C+, whose record names the same series; recall = the share of paired record rounds of that series that the rule labels the same.

| Rule | Frame | Window (days) | Series A precision (%) | Series A recall (%) | Series B precision (%) | Series B recall (%) | Record A or B rounds with a Form D round (%) |
|---|---|---|---|---|---|---|---|
| R1 | B2B software | 180 | 55.9 | 62.9 | 59.2 | 48.4 | 13.3 |
| R2 | B2B software | 180 | 53.9 | 55.0 | 49.4 | 44.1 | 13.3 |
| R1 | all frame | 180 | 49.4 | 57.3 | 56.0 | 42.7 | 12.8 |
| R2 | all frame | 180 | 47.2 | 50.5 | 47.5 | 36.6 | 12.8 |
| R1 | B2B software | 90 | 53.1 | 62.1 | 57.4 | 47.6 | 11.3 |
| R2 | B2B software | 90 | 51.9 | 55.6 | 51.4 | 46.3 | 11.3 |
| R1 | all frame | 90 | 48.0 | 58.2 | 54.9 | 43.9 | 11.0 |
| R2 | all frame | 90 | 46.4 | 52.2 | 48.9 | 38.6 | 11.0 |

Rule R1 against record types, B2B software, 180 days:

| Record type | Records 2015-2024 | Paired with a Form D round | Labelled Series A | Labelled Series B | No label |
|---|---|---|---|---|---|
| seed | 4,780 | 365 | 59 | 0 | 306 |
| A | 1,174 | 151 | 95 | <50 | <50 |
| B | 657 | 93 | <50 | <50 | <50 |
| C+ | 541 | 94 | <50 | <50 | 75 |
| other | 2,855 | 362 | 63 | <50 | 281 |

R1 clears the pre-registered 50% precision bar for both series (55.9% and 59.2%), so results are written as Series A and Series B. Most R1 Series A rounds that the record calls something else are record seed rounds: "Series A" here is best read as the first large priced round. R2's Series B precision is 49.4%, under the bar, so R2's Series B rows are rounds of Series B size. Only 13.3% of record Series A or B rounds have a matched Form D round within 180 days, and the check covers last rounds only, which over-represents companies that stopped raising.

## 5. Measuring hiring

- **Quarters.** q runs from 0 (2015 Q1, quarter end 2015-03-31) to 44 (2026 Q1), the last complete quarter. Base quarter b = the last quarter end strictly before the round's first Form D filing. Form D is due within 15 days of the first sale, so b comes before the money.
- **Windows.** After = from b to b+4 (12 months); before = from b-4 to b. Events are dated 2016-04-01 to 2024-12-31, so both windows lie inside the history and the latest ends in 2025 Q3.
- **Visible team.** Rebuilt headcount at b of 5 to 500 people.
- **Function groups.** From the shared headcount history's function labels, interns (the intern level) left out of every group and founders reported apart:

| Group | Title functions in the taxonomy |
|---|---|
| Engineering | software engineering, other engineering, growth engineering |
| Product | product management |
| Data and analytics | data analysts and scientists, analytics, BI |
| Sales | SDR and BDR, account executives, account management, sales (general), partnerships and business development, sales engineering, revenue operations, enablement, GTM engineering, other GTM |
| Marketing | marketing, demand generation, growth marketing, marketing operations |
| Customer success and support | customer success, customer support |
| Recruiting and HR | recruiting and HR |
| Operations | business operations |
| Other and unclassified | consulting, students, other titles, missing titles |

- **Net adds.** For event i and group g, after = H_g(b+4) - H_g(b) and before = H_g(b) - H_g(b-4), where H is the headcount of the group. Total net adds = the sum over the nine groups. A person with titles in two groups on the same date counts in both.
- **Pooled share.** S_g = the sum over events of net adds in g divided by the sum over events of total net adds. It is the share of all net new people in the group of companies, equal to the ratio of mean net adds.
- **Intervals.** 1,000 bootstrap resamples of companies (a company with both a Series A and a Series B event is one unit), percentile 95% intervals. Each table draws from its own fixed random stream, seeded from the table's name, so a difference within a table uses the same resamples as its parts.

## 6. Comparisons

- **Before against after:** S_g(after) - S_g(before) on the same events.
- **Series B against Series A:** the difference in pooled shares, with companies resampled jointly.
- **Raised again:** events dated up to 2023-09-30, so 36 months of later filings fit inside the extract (Form D filings to 2026-09-30). Raised again = a later Form D round of the same company, of any security type or amount, dated within 36 calendar months. Raw difference in pooled shares, and a raise-year standardized difference: within bins 2016-2017, 2018-2019, 2020-2021 and 2022-2023, raised again minus not, weighted by each bin's share of events (bins with fewer than 10 events in either group dropped; none was).
- **Raise period:** Series A raises dated 2016-2019 against 2022-2023, and the mix by raise-year bin.
- **What is in "other and unclassified":** the person-level roles at the event companies were recounted with the shared headcount history's rule, and titles in the group were split by an ordered set of title patterns (Decision D16 in the study's decision log).

## 7. Results

### 7.1 Mix of net adds after the round

| Group | Series A share (%) | 95% interval | Series A net adds | Series A mean per company | Series A companies with a positive net add (%) | Series B share (%) | 95% interval | Series B mean per company |
|---|---|---|---|---|---|---|---|---|
| Engineering | 21.5 | 19.2 to 23.9 | 2011 | 2.9 | 69.2 | 17.7 | 15.0 to 20.7 | 5.5 |
| Product | 3.7 | 3.1 to 4.3 | 344 | 0.5 | 37.1 | 3.1 | 2.4 to 3.9 | 1.0 |
| Data and analytics | 1.8 | 1.3 to 2.3 | 172 | 0.2 | 18.6 | 1.5 | 1.0 to 2.0 | 0.5 |
| Sales | 19.8 | 17.3 to 22.1 | 1846 | 2.6 | 66.1 | 23.1 | 18.8 to 27.3 | 7.2 |
| Marketing | 7.1 | 6.3 to 8.0 | 667 | 1.0 | 46.1 | 7.1 | 6.2 to 8.3 | 2.2 |
| Customer success and support | 7.7 | 6.2 to 10.0 | 717 | 1.0 | 42.1 | 7.0 | 5.5 to 8.6 | 2.2 |
| Recruiting and HR | 2.0 | 1.6 to 2.5 | 189 | 0.3 | 17.5 | 2.6 | 2.1 to 3.3 | 0.8 |
| Operations | 2.6 | 2.0 to 3.3 | 244 | 0.3 | 24.0 | 2.3 | 1.5 to 3.4 | 0.7 |
| Other and unclassified | 33.8 | 30.1 to 38.4 | 3155 | 4.5 | 78.5 | 35.6 | 29.4 to 43.0 | 11.1 |
| Build (engineering, product, data) | 27.0 | 24.3 to 29.7 | 2527 | 3.6 | 75.8 | 22.3 | 19.2 to 25.9 | 6.9 |
| Sales and marketing | 26.9 | 24.5 to 29.3 | 2513 | 3.6 | 71.4 | 30.2 | 25.7 to 34.7 | 9.4 |
| All nine groups | 100.0 | - | 9345 | 13.4 | 87.6 | 100.0 | - | 31.1 |

Founders (left out of the groups) changed by <50 people in total after a Series A. Engineering minus sales after a Series A: +1.8 points (-1.3 to 4.8).

### 7.2 Before against after

| | A before | A after | A change | 95% interval | B before | B after | B change | 95% interval |
|---|---|---|---|---|---|---|---|---|
| Engineering (share, %) | 22.3 | 21.5 | -0.8 | -3.0 to 1.5 | 19.9 | 17.7 | -2.2 | -4.8 to 0.5 |
| Product (share, %) | 2.8 | 3.7 | +0.9 | -0.1 to 1.8 | 3.2 | 3.1 | -0.1 | -1.1 to 0.9 |
| Data and analytics (share, %) | 2.2 | 1.8 | -0.3 | -1.2 to 0.4 | 1.8 | 1.5 | -0.3 | -1.1 to 0.5 |
| Sales (share, %) | 18.4 | 19.8 | +1.4 | -1.5 to 4.2 | 23.8 | 23.1 | -0.7 | -3.9 to 2.7 |
| Marketing (share, %) | 5.8 | 7.1 | +1.4 | -0.2 to 2.9 | 5.5 | 7.1 | +1.6 | 0.2 to 3.1 |
| Customer success and support (share, %) | 7.6 | 7.7 | 0.0 | -1.4 to 1.6 | 8.3 | 7.0 | -1.4 | -2.9 to 0.3 |
| Recruiting and HR (share, %) | 1.1 | 2.0 | +0.9 | 0.2 to 1.5 | 1.9 | 2.6 | +0.7 | -0.1 to 1.4 |
| Operations (share, %) | 2.2 | 2.6 | +0.4 | -0.4 to 1.1 | 2.4 | 2.3 | -0.1 | -0.9 to 0.7 |
| Other and unclassified (share, %) | 37.6 | 33.8 | -3.8 | -6.6 to -1.1 | 33.1 | 35.6 | +2.5 | -1.4 to 5.7 |
| Build (engineering, product, data) (share, %) | 27.3 | 27.0 | -0.2 | -2.7 to 2.3 | 24.9 | 22.3 | -2.6 | -5.4 to 0.7 |
| Sales and marketing (share, %) | 24.1 | 26.9 | +2.8 | -0.4 to 5.8 | 29.3 | 30.2 | +0.9 | -2.3 to 4.4 |
| Mean total net adds per company | 7.5 | 13.4 | +5.9 | 4.8 to 7.2 | 18.9 | 31.1 | +12.2 | 7.7 to 17.1 |

### 7.3 Series B against Series A

| Group | After Series A (%) | After Series B (%) | B minus A (points) | 95% interval |
|---|---|---|---|---|
| Engineering | 21.5 | 17.7 | -3.8 | -6.3 to -1.0 |
| Product | 3.7 | 3.1 | -0.6 | -1.4 to 0.2 |
| Data and analytics | 1.8 | 1.5 | -0.4 | -1.0 to 0.3 |
| Sales | 19.8 | 23.1 | +3.3 | 0.3 to 7.4 |
| Marketing | 7.1 | 7.1 | 0.0 | -1.3 to 1.3 |
| Customer success and support | 7.7 | 7.0 | -0.7 | -2.9 to 1.3 |
| Recruiting and HR | 2.0 | 2.6 | +0.6 | -0.1 to 1.3 |
| Operations | 2.6 | 2.3 | -0.3 | -1.3 to 0.7 |
| Other and unclassified | 33.8 | 35.6 | +1.8 | -2.6 to 5.8 |
| Build (engineering, product, data) | 27.0 | 22.3 | -4.7 | -7.6 to -1.6 |
| Sales and marketing | 26.9 | 30.2 | +3.3 | 0.2 to 7.6 |

Within size bands at the base quarter the gap in the sales and marketing share is +3.1 points at 5-24 people (-2.7 to 8.9) and +2.2 at 25-99 (-2.5 to 7.0).

### 7.4 Raised again within 36 months

Series A: 318 of 628 events (50.6%) had a later Form D round within 36 months, a median 17.7 months later; 39.3% had a later share sale of $4M or more.

Series B: 125 of 249 events (50.2%) had a later Form D round within 36 months, a median 18.7 months later; 43.4% had a later share sale of $4M or more.

| Series | | Raised again | Did not | Raw difference | 95% interval | Standardized by raise year | 95% interval |
|---|---|---|---|---|---|---|---|
| Series A | Engineering (share, %) | 20.2 | 22.7 | -2.6 | -7.4 to 1.9 | -3.5 | -8.3 to 1.0 |
| Series A | Product (share, %) | 3.7 | 3.8 | -0.1 | -1.5 to 1.1 | -0.0 | -1.4 to 1.3 |
| Series A | Data and analytics (share, %) | 2.0 | 1.6 | +0.4 | -0.7 to 1.7 | +0.6 | -0.5 to 2.0 |
| Series A | Sales (share, %) | 20.3 | 18.2 | +2.1 | -2.9 to 7.3 | +3.1 | -1.0 to 8.0 |
| Series A | Marketing (share, %) | 6.3 | 8.9 | -2.6 | -4.5 to -0.7 | -2.3 | -4.2 to -0.3 |
| Series A | Customer success and support (share, %) | 7.2 | 9.1 | -1.9 | -7.5 to 2.2 | -0.9 | -5.6 to 3.3 |
| Series A | Recruiting and HR (share, %) | 2.0 | 2.1 | -0.1 | -1.0 to 0.8 | +0.1 | -0.9 to 1.2 |
| Series A | Operations (share, %) | 2.6 | 2.6 | 0.0 | -1.3 to 1.4 | +0.5 | -0.9 to 2.0 |
| Series A | Other and unclassified (share, %) | 35.6 | 30.9 | +4.7 | -1.8 to 12.8 | +2.4 | -3.1 to 7.5 |
| Series A | Build (engineering, product, data) (share, %) | 25.9 | 28.1 | -2.2 | -7.7 to 3.3 | -2.9 | -8.3 to 2.2 |
| Series A | Sales and marketing (share, %) | 26.6 | 27.1 | -0.5 | -5.4 to 4.6 | +0.8 | -3.6 to 5.7 |
| Series A | Mean total net adds per company | 17.5 | 10.5 | +7.1 | 3.6 to 11.2 | - | - |
| Series B | Engineering (share, %) | 19.1 | 16.1 | +3.0 | -3.3 to 8.2 | +4.8 | -2.5 to 12.1 |
| Series B | Product (share, %) | 3.0 | 3.3 | -0.3 | -2.4 to 1.2 | -1.2 | -4.2 to 0.5 |
| Series B | Data and analytics (share, %) | 1.5 | 1.6 | -0.0 | -1.1 to 1.0 | -0.0 | -1.3 to 1.9 |
| Series B | Sales (share, %) | 23.4 | 22.3 | +1.1 | -8.5 to 9.9 | -3.6 | -18.9 to 4.5 |
| Series B | Marketing (share, %) | 6.2 | 8.0 | -1.7 | -4.2 to 0.4 | -1.7 | -6.7 to 0.8 |
| Series B | Customer success and support (share, %) | 7.6 | 6.3 | +1.3 | -2.3 to 4.2 | +1.3 | -2.8 to 5.1 |
| Series B | Recruiting and HR (share, %) | 2.6 | 2.6 | +0.1 | -1.3 to 1.3 | +0.7 | -1.3 to 3.1 |
| Series B | Operations (share, %) | 2.9 | 1.5 | +1.4 | -0.2 to 3.4 | +0.9 | -1.0 to 2.9 |
| Series B | Other and unclassified (share, %) | 33.6 | 38.4 | -4.8 | -20.0 to 9.0 | -1.2 | -8.8 to 16.1 |
| Series B | Build (engineering, product, data) (share, %) | 23.6 | 21.0 | +2.7 | -5.2 to 8.7 | +3.5 | -5.3 to 11.5 |
| Series B | Sales and marketing (share, %) | 29.7 | 30.3 | -0.6 | -10.9 to 8.6 | -5.2 | -23.6 to 3.3 |
| Series B | Mean total net adds per company | 37.4 | 28.0 | +9.4 | -5.5 to 21.7 | - | - |

Events per raise-year bin, raised again and not (Series A): 2016-2017: 61 and <50; 2018-2019: 84 and 73; 2020-2021: 103 and 100; 2022-2023: 70 and 99.

### 7.5 Raise period

| Series | Raised | Events | Sales and marketing share (%) | 95% interval | Mean net adds per company | 95% interval (post-hoc) |
|---|---|---|---|---|---|---|
| Series A | 2016-2017 | 99 | 19.0 | 13.3 to 25.9 | 18.8 | 12.2 to 28.5 |
| Series A | 2018-2019 | 157 | 26.8 | 20.5 to 32.2 | 12.4 | 10.4 to 14.8 |
| Series A | 2020-2021 | 203 | 31.0 | 28.0 to 34.0 | 17.1 | 14.2 to 21.0 |
| Series A | 2022-2023 | 191 | 26.9 | 23.1 to 31.2 | 8.8 | 7.0 to 10.7 |
| Series A | 2024 | <50 | - | - | - | - |
| Series B | 2016-2017 | <50 | - | - | - | - |
| Series B | 2018-2019 | 63 | 29.7 | 23.1 to 35.7 | 36.7 | 26.0 to 49.0 |
| Series B | 2020-2021 | 82 | 31.5 | 27.7 to 35.4 | 39.1 | 30.8 to 49.1 |
| Series B | 2022-2023 | 66 | 28.7 | 10.8 to 48.2 | 12.6 | 6.8 to 18.8 |
| Series B | 2024 | <50 | - | - | - | - |

Series A sales and marketing share, 2022-2023 minus 2016-2019: +3.9 points (-2.6 to 10.5). Groups under 50 events are blank.

### 7.6 Team size and the timing of the shift

| Events | n | Sales and marketing share, two years before (%) | Year before (%) | Year after (%) | Change before the raise | Change across the raise | Difference |
|---|---|---|---|---|---|---|---|
| All Series A events with two years of history (pre-registered check k) | 641 | 21.8 | 24.7 | 28.6 | +2.9 (-0.3 to 6.1) | +3.9 (0.6 to 7.2) | +1.0 (-5.1 to 6.8) |
| 5-24 people at the base quarter (post-hoc) | 499 | 17.7 | 19.7 | 28.8 | +2.0 (-2.3 to 6.3) | +9.0 (5.7 to 12.5) | +7.0 (0.5 to 14.3) |
| 25-99 people at the base quarter (post-hoc) | 136 | 26.3 | 30.7 | 29.8 | +4.4 (-0.9 to 9.9) | -0.9 (-6.8 to 4.9) | -5.3 (-15.1 to 4.2) |

### 7.7 What is in "other and unclassified"

The recount of person-level roles at the 783 event companies reproduces the shared company-quarter table by function in 252,067 of 252,067 company-quarter-function cells; the split below adds up to the group in all 26,558 company-quarters.

| Part of other and unclassified | After Series A (% of all net adds) | 95% interval | After Series B (%) | 95% interval |
|---|---|---|---|---|
| Non-staff listings (board, advisors, investors, gig and freelance work) | 9.6 | 6.2 to 14.8 | 9.8 | 3.4 to 18.9 |
| Design | 2.8 | 2.3 to 3.4 | 2.6 | 1.9 to 3.2 |
| Project, program and implementation | 2.8 | 2.2 to 3.5 | 3.8 | 3.0 to 4.7 |
| Technical roles the taxonomy misses | 3.0 | 2.3 to 3.8 | 3.7 | 2.7 to 5.0 |
| Finance, legal, people operations and administration | 4.3 | 3.6 to 4.9 | 5.4 | 4.6 to 6.2 |
| Executives and leaders with no function named | 6.5 | 5.6 to 7.5 | 5.9 | 4.6 to 7.2 |
| Other titles | 4.7 | 3.9 to 5.7 | 4.4 | 3.2 to 5.5 |

Unclassified titles held by 50 or more of the people who joined in the 12 months after a Series A: board member (138); personal shopper (117); investor (83); board observer (78); advisor (74); shopper (61); product designer (55). 4,672 people with an unclassified title joined in those windows.

## 8. Robustness

| Version | Series A events | Build share after A (%) | Sales and marketing share after A (%) | Change across the A (points) | Raised again minus not, standardized (points) | Sales and marketing share after B (%) |
|---|---|---|---|---|---|---|
| Headline | 699 | 27.0 | 26.9 (24.5 to 29.3) | +2.8 (-0.4 to 5.8) | +0.8 (-3.6 to 5.7) | 30.2 |
| (a) Series rule R2 | 631 | 28.7 | 27.2 (24.8 to 29.6) | +2.8 (-0.9 to 6.5) | +1.7 (-3.5 to 7.2) | 29.4 |
| (b) Record names the same series within 180 days | 82 | 40.0 | 20.5 (10.0 to 28.1) | -9.7 (-23.4 to 3.1) | - | - |
| (c) Events dated up to 2023 (late reporting) | 650 | 26.8 | 26.8 (24.3 to 29.6) | +2.8 (-0.9 to 6.2) | +0.8 (-3.6 to 5.8) | 30.1 |
| (d) Broader software and internet labels added | 933 | 26.3 | 26.0 (23.7 to 28.2) | +2.0 (-1.4 to 4.7) | -0.8 (-4.8 to 3.8) | 29.2 |
| (e) 5-24 people at the base quarter | 541 | 29.9 | 27.7 (25.3 to 30.0) | +8.4 (5.1 to 11.2) | -1.9 (-7.0 to 2.7) | 30.7 |
| (e) 25-99 people | 151 | 25.5 | 29.1 (25.1 to 32.6) | -2.2 (-8.1 to 4.0) | +6.4 (-24.8 to 50.2) | 31.3 |
| (e) 100-500 people | <50 | - | - | - | - | - |
| (e) Net adds winsorized at the 1st and 99th percentiles | 699 | 28.7 | 27.9 (26.1 to 29.7) | +4.1 (1.6 to 6.9) | +0.1 (-4.0 to 4.0) | 31.3 |
| (f) Classified groups only | 699 | 40.8 | 40.6 (37.7 to 43.5) | +1.9 (-2.6 to 6.1) | +2.5 (-4.3 to 8.8) | 46.9 |
| (h) Rebuilt headcount 0.7-1.3 times the record's count | 297 | 29.7 | 28.4 (25.3 to 31.3) | +3.2 (-1.3 to 8.0) | +0.3 (-6.0 to 6.3) | 31.0 |
| (i) Raises from 2015 Q2 added (after window only) | 762 | 26.7 | 27.6 (25.2 to 29.9) | - | - | 30.8 |
| (j) Non-staff listings left out | 699 | 29.9 | 29.7 (27.3 to 32.0) | +2.3 (-1.2 to 5.7) | +1.8 (-3.1 to 7.4) | 33.5 |

- Raised again, (g1) later round must be a share sale of $4M or more: sales and marketing difference +1.5 points (-3.1 to 6.3); difference in mean net adds +10.0 people.

- Raised again, (g2) standardized by raise year and size band: sales and marketing difference -0.0 points (-5.6 to 7.5); difference in mean net adds +7.1 people.

Blank cells rest on fewer than 50 companies or 50 people. In the 100-500 band and the record-confirmed Series B set there are fewer than 50 events.

## 9. Pre-registered hypotheses and verdicts

| Hypothesis | Result | Verdict |
|---|---|---|
| H1. Engineering is the largest classified group after a Series A (20-35% of net adds), ahead of sales; sales and marketing 15-30% | Engineering 21.5%, sales 19.8%, engineering minus sales +1.8 (-1.3 to 4.8); sales and marketing 26.9% | Direction: wrong (the lead over sales has an interval that includes 0). Level: holds. |
| H2. The sales and marketing share is at least 3 points higher after a Series A than before; engineering's share is lower | Change +2.8 (-0.4 to 5.8); engineering -0.8 (-3.0 to 1.5) | wrong; engineering part: wrong |
| H3. The sales and marketing share is higher after a Series B than after a Series A | B minus A +3.3 (0.2 to 7.6) | holds |
| H4. Companies that raised again within 36 months put 2-8 points more of their post-A net adds into sales and marketing, and added more people | Standardized difference +0.8 (-3.6 to 5.7); net adds +7.1 (3.6 to 11.2) | Mix: wrong. Volume: holds. |
| H5. Series A raises in 2022-2023 put a smaller share into sales and marketing than raises in 2016-2019 | 26.9% against 23.0% (+3.9, -2.6 to 10.5) | wrong |

## 10. Verification

An independent re-analysis, written from the study plan alone without seeing our code or outputs, recomputed the headline numbers. It received the hypothesis, the plan, Decisions D1 to D19 and the same data extracts, wrote its own code, and returned 107 numbers: 61 counts and point estimates and 46 bounds of 95% intervals. The match rules were fixed before its numbers were read (Decision D20). All 107 match. All 61 counts and point estimates equal the study's unrounded values to six decimals. 34 of the 46 interval bounds are identical, because the re-analysis named those random streams as the study did; the other 12 come from streams it named itself and differ from the study's by at most 0.17 of the allowed tolerance (a quarter of the study's interval width). The re-analysis also rebuilt 50 of the point estimates along a second code path (no mismatches), reproduced the series-rule check exactly and found the same 783 event companies.

Where the plan left room, the re-analysis chose as follows; since its point estimates equal the study's, the study's choices had the same effect. Founder-level rows under other functions (mostly engineering, probably founding engineers) stay in their function group; dropping them would move the engineering share after a Series A from 21.5% to 21.6% and leave the sales and marketing share at 26.9%. The headcount table counts a person with two titles in one group, or at two levels, twice, and leaves out founders and interns; at the base quarter the nine groups add up to 94.6% of the visible team. Raise-year weights in the standardized comparison are recomputed in every bootstrap draw. The period difference is later minus earlier.

## 11. Alternative explanations considered

1. **Rounds labelled as Series A are often large seed rounds.** Under rule R2 the sales and marketing share after a Series A is 27.2%. Among the rounds whose company record names the same series, it is 20.5% (build 40.0%), on only 82 events with a wide interval, so a cleaner set of true Series A rounds may tilt more toward building. Named as a limitation.
2. **Staff abroad are missing.** At companies whose rebuilt headcount is close to the employee count on their record, the shares are build 29.7% and sales and marketing 28.4%, close to the headline. Engineering hired abroad is still a likely downward bias on engineering's share.
3. **Board seats, advisors and gig listings counted as hires.** Without them, build 29.9% and sales and marketing 29.7%: building and selling stay level.
4. **A few large companies drive the pooled shares.** Winsorized net adds give a sales and marketing share of 27.9%; by size band at the base quarter, 27.7% (5-24 people) and 29.1% (25-99).
5. **Late-added jobs.** Events dated up to 2023 (windows end by 2024 Q4): 26.8%.
6. **The industry set.** Adding the broader software and internet labels: 26.0%.
7. **Unclassified titles.** Over classified groups only, build 40.8% and sales and marketing 40.6%: the same balance.
8. **The before-after change is a trend, not the round.** Over the two years before a Series A the sales and marketing share was already rising (+2.9 points), about as fast as across the raise (+3.9); the difference is +1.0 (-5.1 to 6.8). In the post-hoc split by size, teams of 5-24 people show a larger shift across the raise than before it (+7.0, 0.5 to 14.3). The design is an event study without a comparison group, so this is timing, not cause.
9. **Companies that raised again were better companies.** The comparison holds the raise year fixed (and, in g2, size), but not traction, revenue or investor quality. Form D also misses rounds that were not filed, which puts some companies that raised again in the other group and pulls the difference toward zero.
10. **The 2021 boom.** Mixes and volumes are shown by raise-year bin, and the raised-again comparison is standardized by raise year.
11. **The Series B gap is company size.** Series B companies are larger at the base quarter (median 38 people against 14). Within size bands the gap in the sales and marketing share is about as large as overall (+3.1 points at 5-24 people, +2.2 at 25-99, against +3.3) but not clear, so size does not explain it away; the bands are too small to pin it down.

## 12. Limitations

- **Form D coverage.** Form D covers US issuers only, filing is not perfectly enforced, and many small raises, SAFEs and convertible notes are not filed. Name matching misses renamed companies and companies whose legal name differs from the name on their company record; 4.0% of US B2B software records have a matched filing. The sample is the B2B software companies whose filings were matched, not all companies that raised.
- **Series labels are inferred.** R1 agrees with the round type on company records 55.9% of the time for Series A and 59.2% for Series B, on a check that covers last rounds only.
- **Six countries.** Ken internal data holds people whose current country is the US, UK, Canada, Australia, Ireland or New Zealand. Staff elsewhere are missing; for US software companies that hire engineers abroad, engineering's share is understated.
- **Self-reported profiles.** People add new jobs months late, leave short jobs off and sometimes edit an old title instead of adding a new one when promoted. Only currently active profiles exist, so older years undercount; comparisons across years are safer than levels.
- **Net adds, not hires.** Headcount is a stock: a net add can be several hires minus departures, and the data cannot separate them here.
- **A third of titles are unclassified** (33.8% of net adds after a Series A), including non-staff listings such as board seats and gig work.
- **B2B is a proxy.** Industry labels are self-chosen and current; product-software labels include some consumer apps.
- **Visible team.** Events need 5 to 500 people in the data at the base quarter, which drops companies whose staff are mostly abroad or missing.
- **No before window for 2015 raises.** The history starts at the 2015 Q1 quarter end, so the main sample starts with raises dated 2016-04-01; robustness (i) adds 2015 Q2 to 2016 Q1 raises to the after-window mix.
- **Correlation.** Every comparison is descriptive or an event study around a dated filing. Nothing here shows what the money caused.

## 13. Privacy

Only aggregate tables left Ken's servers. Every published statistic rests on at least 50 companies; counts of 1 to 49 companies, events or people are shown as "<50", and shares are suppressed when the group's headcount at the end of the window, summed over companies, is under 50 people. Because subsets overlap, exact people sums are published for the main sample only and means in subset rows are rounded (Decision D19). Titles are listed only when 50 or more people hold them. No person or company is named.

## 14. Sources and reuse

- SEC Form D quarterly data sets (sec.gov), 2008 Q1 to 2026 Q3, through the shared funding events extract (built for the first sales hire study).
- Ken internal data through the shared headcount history (built for the SDR study; build of 2026-10-17), with the shared sales title taxonomy (built for the GTM engineer study).
- The GTM-to-engineering study: the product-software label set, the ratio benchmarks cited in the hypothesis and the list of non-staff title words reused in Decision D16.
- The Ken Research study *What does a normal B2B SaaS pricing page look like?*: the round-type labels on company records.
- Carta, published round-size benchmarks for 2021-2024 (carta.com/data), for the series thresholds.

## 15. Decisions

Every judgment call is logged with its reason in the study's decision log (Decisions D1 onward).