Recent-launch lists drew 4 times the positive reply rate of firmographic lists in B2B cold email
We compared positive replies per send across cold email lists from outbound programs run for 416 B2B companies (5.59M sends), and lists built on a recent launch or a named event beat plain firmographic lists while job-post hiring and customer-lookalike lists fell below them in both periods studied.
Summary
In outbound email programs run for 416 B2B companies between 2022 and 2026 (5.59M sends), lists of recently launched companies drew 10.4 positive replies per 1,000 sends (12 campaigns, 10 clients, 25,496 sends, 265 positives), 4.0 times the 2.6 per 1,000 of plain firmographic lists (483 campaigns, 167 clients, 2,103,450 sends, 5,446 positives), and did better than the same clients' own firmographic lists for 5 of 7 clients. Lists of a named event's attendees or exhibitors drew 13.4 per 1,000 (8 campaigns, 4 clients, 28,565 sends, 382 positives), a directional result because one client sent most of those emails. Contacts who were new in their role also beat firmographic lists in both periods studied, with one large program sending most of that volume each time. Job-post hiring lists and lookalikes of current customers ran below firmographic lists in both periods, third-party intent data plays ran below them, and funding and technographic lists showed no consistent lift. The two periods used different sending platforms and different definitions of a positive reply, so rates should be compared only within a period.
Findings
- Lists of recently launched companies drew 10.4 positive replies per 1,000 sends in 2022 to 2025 (12 campaigns, 10 clients, 25,496 sends, 265 positives), 4.0 times the 2.6 per 1,000 of plain firmographic lists (483 campaigns, 167 clients, 2,103,450 sends, 5,446 positives).
- Lists of a named event's attendees or exhibitors drew 13.4 per 1,000 (8 campaigns, 4 clients, 28,565 sends, 382 positives), 5.2 times the baseline, a directional result because one client sent most of those emails.
- Contacts new in their role beat firmographic lists in both periods, at 1.3 times in 2022 to 2025 (5 campaigns, 5 clients, 29,435 sends, 99 positives) and 2.9 times in 2025 to 2026 (22 plays, 19 workspaces, 70,423 sends, 321 positives), with one program sending most of the volume each time.
- Job-post hiring lists ran below firmographic lists in both periods, at 0.4 times in 2022 to 2025 (12 campaigns, 10 clients, 23,017 sends, 24 positives) and 0.6 times in 2025 to 2026 (19 plays, 13 workspaces, 57,458 sends, 56 positives).
- Third-party intent data plays ran at 0.6 times the baseline (8 plays, 8 workspaces, 12,157 sends, 11 positives), and none of the 8 workspaces did better on them than on its own firmographic plays.
- Funding lists showed no consistent lift, at 0.6 times the baseline in 2022 to 2025 (44 campaigns, 33 clients, 121,591 sends, 186 positives) and 1.2 times in 2025 to 2026 (12 plays, 9 workspaces, 44,982 sends, 87 positives).
- Technographic lists ran at 0.5 times the baseline (63 campaigns, 30 clients, 294,400 sends, 411 positives) and at 1.0 times the same clients' own firmographic lists (22 clients; 50 and 55 campaigns; 265,435 and 185,925 sends; 379 and 274 positives).
- Website visitor lists could not be measured, because one company ran them in each period.
How to read these numbers. Each period is measured against its own baseline: firmographic lists, defined only by industry, size, location, revenue or job title. The two periods ran on different sending platforms and counted positive replies differently, so compare rates only within a period. Rows built on a handful of clients, and rows in which one client sent more than half of the emails (marked in the tables), are directional.
Triggers in 2022 to 2025
Source: Carroway Research analysis of 674 labeled campaigns with 500 or more sends, run for 189 B2B companies on the first of two sending platforms. Counts per row are in triggers-2022-2025.csv.
In the first period, firmographic lists drew 2.6 positive replies per 1,000 sends (483 campaigns, 167 clients, 2,103,450 sends, 5,446 positives). Two kinds of list stood far above that. Recent-launch lists covered companies that launched recently, were listed on a product-launch site or an AI tools directory, or were founded within the last year. Event lists covered the attendees or exhibitors of a named conference, trade show or expo. Both are tied to a dated moment at the company, a launch or a show.
The two results differ in how far they reach. The launch row spans 10 clients, and no single client sent most of it. The event row spans 4 clients, and one of them sent most of its emails, so the event rate mostly describes one program.
New in role lists, built on people recently hired into the target role or past contacts who had moved to a new company, ran above the baseline at 1.3 times (5 campaigns, 5 clients, 29,435 sends, 99 positives), again with one client sending most of the volume. Every other trigger in this period ran below the baseline, from ad or sponsorship spend down to job-post hiring and lookalikes of current customers. The mixed group of other list sources, which holds client-supplied lists, re-sends to earlier prospects, website visitors, social followers and similar lists, sat close to the baseline.
| Trigger | Campaigns | Clients | Sends | Positives | Positive per send (%) | Positives per 1000 sends | Lift vs baseline | Single client majority |
|---|---|---|---|---|---|---|---|---|
| Event attendee or exhibitor list | 8 | 4 | 28,565 | 382 | 1.3 | 13.4 | 5.2 | yes |
| Recent launch | 12 | 10 | 25,496 | 265 | 1.0 | 10.4 | 4.0 | no |
| New in role | 5 | 5 | 29,435 | 99 | 0.3 | 3.4 | 1.3 | yes |
| Other list sources (mixed) | 31 | 21 | 190,519 | 558 | 0.3 | 2.9 | 1.1 | no |
| Firmographic list (baseline) | 483 | 167 | 2,103,450 | 5,446 | 0.3 | 2.6 | 1.0 | no |
| Ad or sponsorship spend | 11 | 8 | 60,201 | 130 | 0.2 | 2.2 | 0.8 | no |
| Funding | 44 | 33 | 121,591 | 186 | 0.2 | 1.5 | 0.6 | no |
| Technographic | 63 | 30 | 294,400 | 411 | 0.1 | 1.4 | 0.5 | no |
| Hiring (job posts) | 12 | 10 | 23,017 | 24 | 0.1 | 1.0 | 0.4 | no |
| Lookalike of current customers | 5 | 4 | 20,008 | 19 | 0.1 | 0.9 | 0.4 | no |
Triggers in 2025 to 2026
Source: Carroway Research analysis of 331 plays with 500 or more sends in 47 client workspaces on the second sending platform, read on 22 September 2026. Counts per row are in triggers-2025-2026.csv.
On the second platform, targeting was set up as plays, and the platform records each play's type. Firmographic plays drew 1.6 positive replies per 1,000 sends (148 plays, 43 workspaces, 482,443 sends, 767 positives). Newly hired executive plays drew 4.6 per 1,000 (22 plays, 19 workspaces, 70,423 sends, 321 positives), 2.9 times the baseline. One workspace sent most of those emails.
Three more play types ran at 1.7 times the baseline or higher. Plays aimed at people who follow a competitor's company page ran at 2.0 times (6 plays, 5 workspaces, 14,706 sends, 47 positives). Lookalikes of accounts the client named as ideal customers ran at 1.9 times (22 plays, 14 workspaces, 31,921 sends, 94 positives). Custom plays, with targeting written by hand for one client, ran at 1.7 times (64 plays, 25 workspaces, 194,564 sends, 537 positives). Funding plays ran slightly above it (1.2 times; 12 plays, 9 workspaces, 44,982 sends, 87 positives).
Lookalikes of current customers, job-post hiring and third-party intent data ran below the baseline. Plays built on competitors named in job posts and on companies running ads drew one positive reply each (3 plays and 3 workspaces each; 1,960 and 3,316 sends), too few to read. Four play types are left out because fewer than 3 workspaces ran them at 500 or more sends: recent launches, engagement with a social post, website visitors and users of a competitor's product.
| Trigger | Plays | Workspaces | Sends | Positives | Positive per send (%) | Positives per 1000 sends | Lift vs baseline | Single client majority |
|---|---|---|---|---|---|---|---|---|
| Newly hired executive | 22 | 19 | 70,423 | 321 | 0.5 | 4.6 | 2.9 | yes |
| Follower of a competitor's company page | 6 | 5 | 14,706 | 47 | 0.3 | 3.2 | 2.0 | no |
| Lookalike of named target accounts | 22 | 14 | 31,921 | 94 | 0.3 | 2.9 | 1.9 | no |
| Custom play (free-form targeting) | 64 | 25 | 194,564 | 537 | 0.3 | 2.8 | 1.7 | no |
| Funding | 12 | 9 | 44,982 | 87 | 0.2 | 1.9 | 1.2 | no |
| Firmographic list (baseline) | 148 | 43 | 482,443 | 767 | 0.2 | 1.6 | 1.0 | no |
| Lookalike of current customers | 24 | 12 | 40,791 | 46 | 0.1 | 1.1 | 0.7 | no |
| Hiring (job posts) | 19 | 13 | 57,458 | 56 | 0.1 | 1.0 | 0.6 | no |
| Third-party intent data | 8 | 8 | 12,157 | 11 | 0.1 | 0.9 | 0.6 | no |
| Competitor named in job posts | 3 | 3 | 1,960 | 1 | 0.1 | 0.5 | 0.3 | no |
| Running ads | 3 | 3 | 3,316 | 1 | 0.0 | 0.3 | 0.2 | no |
Against the same clients' own lists
Source: Carroway Research. Each row keeps only the clients or workspaces that sent both kinds of list in that period. Clients, campaigns or plays, sends and positives for both sides of every row are in within-client-lift.csv.
A pooled rate mixes the trigger with the clients who chose it. A client with a strong offer tends to lift every list it sends, and a client with a weak offer tends to pull every list down. To separate the two, we kept only the clients that sent both a trigger list and a firmographic list in the same period, and compared their trigger rate with their own firmographic rate.
Recent launches held up best across clients. Seven clients sent both kinds of list. Their launch lists drew 11.2 positive replies per 1,000 (9 campaigns, 16,666 sends, 186 positives) against 3.6 on their own firmographic lists (35 campaigns, 92,927 sends, 331 positives), 3.1 times as many, and 5 of the 7 clients did better on launch lists.
The other large lifts rest on fewer programs. Event lists did better than the client's own firmographic lists for 2 of 4 clients, new in role lists for 2 of 5, and newly hired executive plays for 7 of 17 workspaces. In each case one program sent most of the trigger emails, so the pooled lift largely reflects that program.
Technographic lists matched the same clients' own firmographic lists (1.0 times; 22 clients; 50 and 55 campaigns; 265,435 and 185,925 sends; 379 and 274 positives). Their low pooled rate comes from the clients who used them. Funding lists stayed below the same clients' firmographic lists in 2022 to 2025 (0.5 times; 22 clients; 29 and 108 campaigns; 57,052 and 396,495 sends; 81 and 1,024 positives) and were about even in 2025 to 2026 (1.1 times; 7 workspaces; 9 and 36 plays; 37,626 and 196,434 sends; 86 and 396 positives). Job-post hiring did better than the client's own firmographic lists for 2 of 8 clients and 2 of 12 workspaces, and third-party intent data for none of 8 workspaces.
Method
Source. Outbound email programs run for 416 B2B companies between 2022 and 2026, 5.59M sends in total (5,587,325). 245 of the 416 companies have recorded sends. One outbound email operator ran the programs on two sending platforms in turn, and the history includes the operator's own outbound program. Data were read on 22 September 2026.
Two periods. 2022 to 2025 is the first platform: client sign-ups start in 2022, and recorded campaign send dates run from 3 April 2023 to 26 May 2025 (some records have no dates). 2025 to 2026 is the second platform, which began sending on 6 October 2025. Per-play counts exist for plays created from 2 March 2026.
Units. In 2022 to 2025 the unit is a campaign, one list with one email sequence. In 2025 to 2026 it is a play, one targeting recipe inside a client's workspace. A client can run many of either.
How 2022 to 2025 campaigns were bucketed. Each campaign name follows a fixed pattern of client, companies, people and owner, written by the operator's staff when the campaign was set up. We treated the companies and people parts as the targeting label and applied keyword rules in a fixed order. The first rule that matched set the bucket, and labels that matched no rule form the firmographic baseline. No campaign was labeled by hand. We read all 684 labels with 500 or more sends to check the rules, and added exclusions for keywords that misfired, for example a label describing non-venture-backed companies under funding. Every rule is in definitions.csv.
How 2025 to 2026 plays were bucketed. The platform records a type and a signal for each play when it is created. We used those fields as recorded, with no keyword rules. Custom plays, whose targeting was written by hand, are their own row and were not reclassified.
Positive reply. The two periods count positives differently.
- 2022 to 2025: the positive-reply count stored on each campaign record by the operator's reply-handling tools. Replies in that period were sorted into labels such as possibly interested, interested question, interested but later and interested referral. The exact set of labels behind the stored count is not recorded with the data.
- 2025 to 2026: the number of distinct prospects on a play who sent a reply that was not automated and that the platform's automated classifier labeled possibly interested, question, contact later or manual referral. The platform stores this count. We confirmed the rule by recomputing it from reply labels for the 728 plays with replies. It matched the stored count exactly for 722 plays and came within 2 for the other 6.
Sends. Every email sent, follow-ups included. For 2022 to 2025 campaigns we used the larger of the record's two send counters, which disagree on some records. For 2025 to 2026 plays we used the play's sent count.
Rates and lift. Positive replies per 1,000 sends is positives divided by sends, times 1,000, pooled over every campaign or play in a row. The percentage columns are the same figure per 100 sends. Lift is a row's rate divided by the firmographic baseline rate of the same period.
Same-client comparison. For each trigger row we kept the clients or workspaces that also sent firmographic lists in that period, pooled their trigger sends and their firmographic sends separately, and divided the two rates. The clients_trigger_higher column counts the clients whose own trigger rate beat their own firmographic rate. A row needs at least 3 such clients.
Exclusions.
- 178 first-platform records have no reply counters. 20 of them hold 1,195,753 sends as per-client totals added by a later data sync, with no positive count, and the rest hold no sends. All 178 are excluded.
- Campaigns and plays with fewer than 500 sends are excluded in both periods.
- 10 campaigns with 500 or more sends have no targeting label (13,932 sends) and are excluded.
- A row needs at least 3 clients or workspaces. In 2025 to 2026, four play types fell short (5 plays, 10,257 sends) and three more had no play with 500 or more sends, so they are left out. In 2022 to 2025, website visitor lists came from 1 client and are counted in other list sources.
After exclusions, 2022 to 2025 covers 674 campaigns from 189 clients (2,896,682 sends, 7,520 positives), and 2025 to 2026 covers 331 plays from 47 workspaces (954,721 sends, 1,968 positives).
Limits.
- Small rows are directional, and so are rows in which one client or workspace sent more than half of the emails, marked yes in the single_client_majority column: event lists and new in role lists in 2022 to 2025, and newly hired executive plays in 2025 to 2026. The operator's own outbound program sent most of the emails in that last row.
- The periods differ in sending platform, positive-reply definition, unit and bucketing method. Compare rates only within a period.
- Clients chose their own triggers, offers and copy, so a pooled row mixes the trigger with the clients who used it. The same-client comparison narrows this, but one client's offer and copy can still differ from campaign to campaign. The pooled 2022 to 2025 baseline includes a few clients with unusually high rates, which is one reason to read the same-client comparison alongside it.
- Targeting labels are short descriptions written for internal use. A label can leave out a trigger that was used, and the event rule does not check whether emails went out before or after the event.
- Rates count follow-up emails as sends, and sequence length varies between campaigns.
Data
The five CSV files below hold every number in this note. triggers-2022-2025.csv and triggers-2025-2026.csv give one row per list type with its counts, rates and lift. within-client-lift.csv gives the same-client comparison. definitions.csv lists the rule behind every row, and coverage.csv traces the counts from the full history to the rows analyzed. They are free to reuse with credit to Carroway Research.
- Positive replies by list type, 2022 to 2025 One row per list type on the first sending platform: campaigns, clients, sends, positive replies, positives per 100 and per 1,000 sends, lift against the firmographic baseline, and whether one client sent more than half of the row's emails.
- Positive replies by play type, 2025 to 2026 One row per play type on the second sending platform: plays, workspaces, sends, positive replies, positives per 100 and per 1,000 sends, lift against the firmographic baseline, and whether one workspace sent more than half of the row's emails.
- Trigger lists against the same clients' firmographic lists For each trigger row, only the clients or workspaces that also sent firmographic lists in the same period: campaigns or plays, sends, positives and rates on both kinds of list, the ratio between them, and how many clients did better on the trigger list.
- How each list type was assigned The keyword rules behind each 2022 to 2025 row, in the order they were applied, and the platform play type behind each 2025 to 2026 row.
- Coverage and exclusions Companies, campaigns, plays, sends and positives at each step from the full outbound history to the rows analyzed, with date ranges and the check of the 2025 to 2026 positive-reply definition.
How to cite: Carroway Research. "Recent-launch lists drew 4 times the positive reply rate of firmographic lists in B2B cold email." 23 September 2026. https://carrowayresearch.com/reports/cold-email-triggers