--- title: GrowthPal research and methodology url: https://growthpal.io/research/ last_reviewed: 2026-10-24 source: growthpal.io --- # GrowthPal research and methodology This is the reference layer for GrowthPal's method: how a sending list is built, how a reply is classified, which denominator sits under every published figure, and which questions the data cannot currently answer. It exists so that claims made elsewhere on this site can be checked rather than taken on trust. How the work is done and how it is counted, written so that a figure lifted off this site still means what it meant here. Every page names its denominator and states what it does not know. ## Start here | Page | What it answers | | How we measure | Why one campaign supports four different reply rates, which one GrowthPal publishes, what counts as an interested reply, and why opens and clicks are not tracked at all. | | How lists are built | The two passes, the verification step, the under 2% bounce average it protects, and the market size below which the method does not apply. | | Every published result | The seven documented outcomes and the study, each with the denominator it was measured against and who the work belongs to. | | The methodology block | The short-form definitions table, kept in the glossary beside the terms it defines. | ## The published evidence Every campaign figure on this site traces to one of these. None is a modelled or illustrative number. | Source | What it establishes | | The 412-campaign study | 412 campaigns across 47 companies over 18 months, scored on four dimensions. Campaigns above 0.7 on the composite score averaged 11.4% reply rate against 1.1% below 0.3. That is a total reply rate and is not comparable with the interested-reply figures elsewhere. | | The case studies | Four documented outcomes read from sending-platform dashboards rather than calculated, including the month of 67,934 contacts and 449 interested replies. | | The benchmarks | Interested-reply ratios, sends per inbox, warm-up length and the minimum market size, gathered from the same campaigns. | | The deliverability checklist | The infrastructure rules those campaigns ran on, including authentication and the warm-up ramp. | | How an angle is written | The 10 to 15 angles a test puts into market, and what separates one that works. | ## What this research does not establish Four gaps are carried openly across these pages rather than filled with an estimate. They are listed here so a reader does not have to find them one page at a time. - The denominator behind the study's 11.4% reply rate. It is described as a total reply rate but it is not stated whether it is measured against contacts or against messages. Those two differ by a factor of 2.4 on a documented campaign, so the figure cannot be placed against anything else here until it is pinned down. - The exact date range behind the 67,934, 449 and 83.47% figures. - Whether the sending platform's "positive" label is the same classification as "interested" as defined here. - What counts as an opportunity. The 5,000+ opportunities figure is published without a definition, so none is offered. ## How sources are chosen and checked Three rules govern what may be cited on this site, and they are the reason the industry pages read the way they do. - Read the original, not an article quoting it. A regulation is cited at the regulator, a statute at the statute, a statistic at the body that collected it. A blog post summarising a rule is not a source for that rule. - Record the date it was read. Every citation carries an accessed date, because a live document can change after it is quoted and a reader needs to know which version was seen. - If it cannot be verified, cut it. A claim that loses its source loses its place on the page. Sources live on the page that uses them rather than in a central list, directly under the claim they support. Across the site that is 218 documents from 159 publishers, 179 of them primary sources, each with the date it was last read. A source being cited is not an endorsement of the claim attached to it, and where a source is a company writing about its own product the page says so. Vendor documentation is evidence of what a vendor says, never of whether it is true. ## What a link check on 24 October 2026 found Every cited URL was requested automatically. 156 answered immediately. The rest are worth reporting, because an automated check turns out to be a poor judge of whether a document still exists. | Result | Count | What it actually meant | | 200 OK | 156 | Resolved on the first request. | | 403 Forbidden | 48 | Concentrated on government and publisher domains that refuse automated agents. A sample was re-opened properly and each loaded normally. Not broken. | | 404 Not Found | 6 | Each re-checked individually. Four were false alarms from the same blocking, including the FTC's CAN-SPAM compliance guide, which this site cites on several pages and which is perfectly alive. Two were genuinely gone. | | Connection failed | 4 | Sites that drop automated connections outright. Not broken. | | Other | 6 | Rate limiting, paywall probes and one server error. | The lesson is worth stating plainly, because link-rot audits get published without it: a bare status code is not evidence. Forty-eight of these sources would be reported dead by a crawler and every one of them is alive. The two that had actually gone had to be opened to find out. Two citations returned a genuine 404 and were deleted from the page that used them, along with the entries in the data file behind it, so a rebuild cannot reinstate them: TechTarget's deal registration definition and embedded world's exhibitor directory, both previously on the B2B technology page. Neither was load-bearing, and no replacement URL was substituted, because substituting an unverified link is the error the third rule exists to prevent. One was flagged rather than removed. The US Census Bureau's quarterly retail e-commerce page returns a server error on every attempt. A server error is not evidence that a document is gone, so the citation on the ecommerce page stands and is marked for re-check. If it still fails at the next review it goes. ## How to cite these pages Figures here may be quoted with attribution to GrowthPal and a link to the page carrying the figure. Two conditions make a quotation accurate rather than merely permitted: name the denominator alongside the number, and keep the attribution on the Passionfruit figures, which are work Prabesh Khanal led at a previous agency, published with written permission, and not a GrowthPal client engagement. A plain-text version of every page on this site sits beside it, and /llms.txt indexes the set for language models. ## Frequently asked questions about this research **Is any figure on this site modelled or illustrative?** No. Every campaign figure is read from a sending-platform dashboard or comes from the 412-campaign study. Where a number would have to be derived, such as meetings booked from the 449 interested replies, it is left unstated rather than calculated. **Why publish the gaps rather than fill them?** Because a figure whose denominator or date range is unknown is not improved by guessing one. Naming the gap lets a reader judge how much weight the figure carries, which an estimate would quietly remove. **Can these figures be compared with another agency's?** Only where that agency states its denominator. How we measure shows a single campaign producing rates from 0.28% to 1.41%, so two published numbers can differ by five times while describing identical performance. Sources for this page - FTC, CAN-SPAM Act: A Compliance Guide for Business (primary source). Accessed 2026-10-24. - Google, Email sender guidelines (primary source). Accessed 2026-10-24. - Yahoo, Sender Best Practices (primary source). Accessed 2026-10-24. Last reviewed 2026-10-24. A plain-text version of this page is at /research.md , and the site index for language models is at /llms.txt . ## Sources - FTC, CAN-SPAM Act: A Compliance Guide for Business (primary source). https://www.ftc.gov/business-guidance/resources/can-spam-act-compliance-guide-business (accessed 2026-10-24) - Google, Email sender guidelines (primary source). https://support.google.com/a/answer/81126 (accessed 2026-10-24) - Yahoo, Sender Best Practices (primary source). https://senders.yahooinc.com/best-practices/ (accessed 2026-10-24) GrowthPal does not publish a price list and does not guarantee results.