EVERY CLAIM RE-READ 12 AUGUST 2026  //  and one agency we previously ranked no longer has a working websiteApply
Every claim re-read 12 August 2026

Cold email agencies for SaaS, and what their numbers actually mean

Six agencies publish SaaS proof. They publish it in six different units - appointments, positive replies, MQLs, pipeline, ARR, percent of an unstated target - and 12 of 15 published figures carry no denominator at all. This page puts every claim in one table with the URL it came from, and measures the two published rates against 81 of our own campaigns. For the wider category rather than the SaaS slice, the best outsourced SDR companies are compared on published price and commercial model.

Every SaaS claim, side by side

Selected SaaS claims from the vendor pages reviewed on 12 August 2026. Each row retains the reported outcome unit and source. These are historical, self-reported case claims; confirm current offers and the underlying records with the provider before buying.

Read the denominator and time period alongside the outcome. Reply rates, meetings, pipeline and revenue describe different stages. Missing context prevents a like-for-like rate comparison; it does not by itself invalidate a reported count. ReplyLead publishes this comparison and is a competing provider.
AgencyClient or claim What they publishUnit Denominator given?
ReplyLeadReply rate, pooled2.55%unique replies / leads contactedyes - 6,038 over 236,927
Reply rate, median campaign2.12%unique replies / leads contactedyes - 81 campaigns
Belkins
belkins.io/case-studies
read 12 August 2026
Autodesk35 appointments in 4 monthsappointmentsno
Autodesk110% of KPI achievedpercent of an unstated targetno
Terrascope56 enterprise opportunities, $2.2M pipelineopportunities and pipeline dollarsno
Healthcare SaaS, unnamed44 qualified meetings in 5.5 monthsmeetingsno
Growth Rhino
growthrhino.com
read 12 August 2026
Sortly, Series A inventory SaaS112+ positive replies, 50 meetings, $225K new ARRpositive replies, meetings, ARRno
Halo AI60+ positive replies in 30 dayspositive repliesno
ShareChest, fintech11.4% reply rate, 47 meetings a month at peakreply rateno - the base is not stated
SalesCaptain
salescaptain.io
read 12 August 2026
Elastic$320,000+ new revenue in 2022revenueno
B2B SaaS, unnamed45% positive reply ratepositive reply rateno - the base is not stated
B2B SaaS, unnamed50+ booked meetingsmeetingsno
Martal GroupB2B SaaS, unnamed, 8 months22,832 emails, 112 replies, 544 leads, 416 MQLs, 66 SQLs, 62 booked meetingsa full funnelpartly - emails and replies are both given
Jedoxtestimonial only, no metricsnoneno
SalesHive
saleshive.com/pricing
read 12 August 2026
Pricing"$0 Setup fees, ever" is the only number on the pricing pagesetup feenot applicable

12 of 15 published figures give no base. "50+ booked meetings" out of how many contacted? "112+ positive replies" from what volume? These are real results and we have no reason to doubt any of them - but as published they cannot be turned into a rate, which means they cannot be compared to each other or to what you would get.

The exception worth crediting. Martal Group publishes both halves: 112 replies from 22,832 emails over eight months. That is a figure a reader can actually compute (0.49 percent of emails sent), and it is the only competitor claim on this page that survives contact with arithmetic. The same case gives 62 booked meetings from 544 leads, or one meeting per 8.8 leads worked.

Named clients are rarer than case studies. Of Belkins' 13 published case studies, 5 are software or SaaS, and 3 of those name the client. Across all 6 live agencies here, 4 of the 15 published figures describe an unnamed client, which removes the last way a reader could verify them independently.

What those rates look like measured

Two of the claims above are rates rather than counts: Growth Rhino publishes 11.4 percent for ShareChest, and SalesCaptain publishes 45 percent for an unnamed SaaS client. Neither states a base, so neither can be checked. What we can do is show what those numbers would mean against a set that does state its base - ours.

The retained 12 August 2026 snapshot covers 115 campaigns with sends; 81 campaigns contacted at least 500 leads and qualify for this comparison. They contain 236,927 contacted leads and 6,038 unique replies: pooled 2.55%, median 2.12%, 25th percentile 1.38% and 75th percentile 2.97%. Automatic replies are included. These are cumulative campaign observations across five client and two internal programmes, not SaaS-only results or a fixed-duration trial. See reply-rate distribution and methodology.

Place a claimed rate against a measured set

Choose one of eight reference rates. Without JavaScript, the default comparison is 11.4%; all cohort definitions and representative statistics remain readable.

Move between the eight reference rates
0 of 81of our campaigns reached it
100%of eligible campaigns below the selected rate
4.5xour pooled reply rate
2.97%75th percentile in the eligible sample
Representative campaign reply-rate percentiles and a selected claimThree bars show P25 1.38%, median 2.12% and P75 2.97% for 81 eligible campaigns. A dashed line shows the selected claim, capped at the 12% chart boundary. The selected value and sample counts are also provided as text.0%3%6%9%12%claimed 11.4%P251.38%Median2.12%P752.97%
Bars show representative campaign percentiles, not individual campaigns or the full range. The dashed claim line is capped at the 12% chart boundary; the numeric result displays the full selected rate. Cross-provider comparability depends on matching definitions and populations.

Comparison limits. A position against our campaign sample does not test a vendor's claim, establish a performance ceiling or forecast your results. Lists, offers, audiences, time periods and reply definitions differ. Ask for each rate's numerator, denominator and automatic-reply treatment before comparing it with this sample.

Check the evidence behind a shortlist

Confirm the provider's current service, contracting entity, source pages and dated client references. A case study describes a selected engagement; ask how its audience, scope and outcome definitions compare with your planned campaign.

A site that is temporarily unavailable does not prove that a company has closed. Verify availability and service status through current public records before drawing a conclusion.

How to check any of it yourself

Four questions, in the order that removes the most uncertainty per minute. None of them requires you to trust this page.

1. What is the base?

For a rate, request the numerator and denominator; for a count, ask for the reporting period and activity volume. A missing public denominator limits what a reader can verify. It does not establish whether the vendor measures campaigns internally.

2. What counts as a reply, a meeting, a qualified lead?

A positive reply requires an explicit classification rule. Ask how automatic replies, referrals, objections and buying interest are treated. Our sample counts unique replies divided by contacted leads and includes automatic replies; it does not measure positive replies, held meetings or closed revenue. See the definitions and limitations.

3. Ask for representative campaign percentiles

In our eligible 81-campaign sample, the 25th percentile is 1.38%, the median 2.12% and the 75th percentile 2.97%. These are positions in a distribution of campaign rates, not calendar quarters or average rates for the top and bottom quarters. Ask for representative results alongside selected case studies.

4. Check the domain still resolves

Check that the current service and its cited case studies remain accessible. A broken link or temporary DNS failure is a reason to verify, not proof that a provider is no longer trading.

The technographic data tool can inspect public domain signals. DNS observations can suggest infrastructure providers; they do not verify service quality, legal status, a complete technology stack or campaign outcomes.

The four questions, what a usable answer looks like, and what it tells you when the answer does not come. None of these requires you to trust this page.
AskA usable answer If they cannot answer
What is the base? "112 replies from 22,832 emails" - both halves, as Martal publishes The public claim lacks enough context for an independent rate comparison
What counts as a reply or a meeting? Separate definitions for replies, positive replies, booked meetings and held meetings, with explicit treatment of automatic replies and no-shows Outcome stages may be defined differently, preventing a fair comparison
What are your representative campaign percentiles? A defined cohort with representative percentiles; our eligible sample has P25 1.38%, median 2.12% and P75 2.97% Selected case studies do not establish the typical outcome
Does the domain resolve? The site loads and the case studies are still on it Current source availability needs confirmation; an outage alone does not prove closure

Method

The vendor claims

Every figure in the table was read from the vendor's own live page on 12 August 2026, and each row names the URL. Nothing is taken from a directory, a listicle or a press release. Where a vendor publishes several SaaS figures we list them all rather than choosing the most flattering or the least.

Our own numbers

The first-party sample comes from the retained 12 August 2026 snapshot: 115 campaigns with sends, of which 81 contacted at least 500 leads. The eligible sample contains 236,927 contacts and 6,038 unique replies, pooled 2.55%, median 2.12%, P25 1.38% and P75 2.97%. Automatic replies are included. The population spans five client and two internal programmes and is not a SaaS-only benchmark. Definitions and limitations.

What we are not claiming

  • That any vendor is exaggerating. We cannot know, and that is the point - an uncheckable figure is uncheckable in both directions.
  • That our rates are a ceiling or forecast. Another audience or offer may produce different results. These observations do not predict a competitor's performance or your campaign.
  • That we are the best choice for SaaS. We have no named SaaS case study to show, which on this page's own criterion counts against us and is stated in the table.
  • That this list is complete. These are selected publicly documented examples. Other providers exist; inclusion is not certification or a performance ranking.

Our standards are on methodology and editorial standards.