Interview With Barn2 Co-Founder Katie Keith - 20i®  🔗 **Source:** [Interview With Barn2 Co-Founder Katie Keith - 20i®](https://www.20i.com/blog/interview-with-barn2-co-founder-katie-keith/) > [!note]+ Document Notes > tags: [wordpress-plugins, remote-work, ai-search-disruption, product-diversification, public-speaking] > author: Katie Keith > source: 20i > reference: https://www.20i.com/blog/interview-with-barn2-co-founder-katie-keith/ > date created: 2026-09-10 > > > [!info] Generated analysis — not the author's words. Sections below are machine-written > > from the captured source. Only text inside `> [!quote]` blocks is verbatim from the author. > > Do not quote or attribute any other line to the named author. > > # 🏷️ Title: Interview with Barn2 co-founder Katie Keith - 20i® > <!-- deep-read:sections start --> ## 🧭 Table of Contents - [📝 Summary](#-summary) - [📖 Main Content](#-main-content) - [🗒️ Notes](#️-notes) - [📚 References](#-references) - [🗣️ Paraphrase](#️-paraphrase) - [📰 Newsletter](#-newsletter) - [✅ Action Items](#-action-items) - [🧪 Hypotheses](#-hypotheses) - [📊 Methodology](#-methodology) - [❓ Questions](#-questions) - [🔮 Future Research](#-future-research) - [💡 Implications](#-implications) - [🧠 Learning Styles Assessment](#-learning-styles-assessment) - [📊 Accessibility Assessment](#-accessibility-assessment) - [🌍 DEIB Assessment](#-deib-assessment) - [✅ Logical Validity](#-logical-validity) - [🛡️ DARVO Analysis](#-darvo-analysis) - [🎯 Conclusion](#-conclusion) - [🔄 Refactor Appendix](#-refactor-appendix--interview-with-barn2-co-founder-katie-keith---20i) ### 🔗 References - [Barn2](https://barn2.com/) — the WordPress plugin company Keith co-founded and runs as CEO. - [documentlibrary.pro](https://documentlibrary.pro) — Barn2's cloud/SaaS spinoff of its Document Library Pro plugin. - [Embedded video](https://www.youtube.com/watch?v=cAkgXq9LJuU) — clip embedded in the article's public-speaking section. ## 📝 Summary - **What happened.** 20i interviewed Katie Keith, co-founder and CEO of the WordPress plugin company Barn2, about the company's pivot from client agency work to products, its fully remote team, a personal time audit that changed how she delegates, AI's effect on WordPress and on Barn2 specifically, the company's expansion into Shopify apps and a SaaS product, its international sales logistics, and her path into public speaking. - **Why it matters.** Keith names a concrete, dated business problem — two years of declining new-plugin sales that she attributes to the shift from Google organic search to AI-mediated search — and a second, separate AI threat: "vibe coding" substituting for small utility plugins outright. Both are stated as first-hand, unverified claims from inside a single company, not industry-wide data. - **Key stakeholders.** Barn2's roughly 18-person remote team (as of the CEO time audit, about a year before this interview); Barn2's plugin customers; the wider population of WordPress plugin businesses exposed to the same organic-to-AI-search shift Keith describes; 20i, a WordPress hosting company, as publisher. - **Risks and tradeoffs.** Every causal claim in the piece is Keith's own attribution, not measured against competitors, her own conversion-rate data, or a counterfactual. The remote-team, AI-search, and public-speaking sections are each built on a single company's or individual's experience generalized into advice. - **Bottom line.** Barn2's response to a two-year sales decline it attributes to AI search — diversify into Shopify apps, a SaaS product, and non-search marketing channels — is the interview's throughline, framed as a direct response to a threat Keith says is not fully recoverable through better content or SEO alone. ## 📖 Main Content Keith frames Barn2's origin as a 2016 pivot away from agency work. She and her husband and business partner, Andy, ran a WordPress agency but found full-time client responsibility unsustainable — clients expected to reach her personally, which didn't scale. In January 2016, Andy stepped away from client work to build plugins full time; the first plugin launched in March 2016 and, in Keith's words, "started getting sales very quickly." The couple had discussed selling WordPress products for years before actually allocating resources to it — the shift happened only once client work was deliberately deprioritized, not through gradual addition alongside it. The remote-team section makes a specific claim about timing: Barn2 chose remote work partly to stay small, and that held "for a few years" until sales growth outpaced what the founders could support alone. Keith says the norm shifted in Barn2's favor — by the time Barn2 needed to hire beyond the founders, remote work had become standard in the WordPress space, removing a constraint that would have existed earlier. She locates herself outside major tech hubs (southwest England, then Mallorca) and treats that as a feature of remote hiring, not a limitation, since location becomes irrelevant to who she can hire. A self-audit of her own time, run about a year before this interview when Barn2 had roughly 18 people, is presented as the direct cause of two structural changes: delegating specific low-level tasks she'd been doing personally, and hiring a marketing coordinator. The audit also surfaced a values point — she found podcasts, events, and personal-brand work personally rewarding despite them lacking direct, measurable sales attribution, and treats permission to enjoy immeasurable work as a burnout-prevention move for CEOs generally, not just her own case. On AI's effect on WordPress, Keith separates two distinct threats that get conflated in "AI eats WordPress" commentary: AI as a website-building substitute, and AI as a plugin-building substitute (vibe coding). On the first, she expects impact but qualifies it heavily — only "a small proportion" of people will build sites directly with AI; the larger group will use AI-assisted website builders that still run on platforms like WordPress, which she frames as WordPress retaining its security and plugin-infrastructure advantage even as the building process changes. She cites her own side project — vibe-coding a site but deliberately choosing WordPress as the platform so she could install existing plugins for contact forms rather than have the AI build that functionality from scratch — as a first-hand example of this dynamic. On the second threat, she draws a line between vibe-coding a "business-critical plugin" (which she calls irresponsible) and vibe-coding "minor features," concluding that small utility plugins face the most exposure to being replaced outright by AI-built code, and names Barn2 as itself owning several plugins in that category. The interview's most concrete, dated business claim concerns search. Keith states that the shift from organic Google search to AI-mediated search has been "disastrous" for Barn2, and specifically attributes two years of declining new-plugin sales to it. Her mechanism: Barn2 ranked highly in Google organic, and while AI tools do recommend Barn2's plugins, an AI recommendation converts far less often than a direct organic visit, because a visitor who lands on Barn2's own site sees trust signals — professionalism cues, calls to action, newsletter signups — that don't carry over into an AI-generated answer. Her stated response is not to abandon content or SEO but to add AI-recommendation optimization alongside them (framed as largely overlapping efforts), combined with channels she says don't depend on search at all: cross-promotion, partnerships, and products that sell through other channels. She's explicit that content and SEO changes alone won't recover the lost sales. That search-driven revenue pressure is presented as the direct cause of Barn2's move into Shopify apps and cloud/SaaS products — not a separate growth ambition. She notes Shopify app installs come overwhelmingly through the official app store rather than organic search, which insulates that revenue line from the same AI-search pressure hitting WordPress plugin sales. Barn2's cloud product, documentlibrary.pro, is described as WordPress-powered under the surface (a SaaS wrapper around the existing Document Library Pro plugin) — reusing the company's existing technical skill set while reaching an audience that doesn't need to already run WordPress. On international growth, Keith frames software distribution as structurally easier than physical goods (no shipping), but names concrete compliance and localization work: handling sales tax per country's legislation (Barn2 built and sells an Easy Digital Downloads plugin to solve this for others), and ensuring software supports translation, special characters, and right-to-left languages — despite which the Barn2 website itself remains English-only and still gets daily international sales. The public-speaking section is advice drawn from her own escalation path: a joint talk first, filmed practice posted publicly for feedback, a ten-minute solo lightning talk before a full-length solo talk at a smaller event. She reports that studying other speakers' mannerisms on YouTube taught her nothing she didn't already do naturally once recording herself, and concludes people can overthink stylistic mannerisms relative to just presenting as themselves. She separately names being "a woman in tech who owns a reasonably well established company" as something that made it easier for organizers to say yes to her, and explicitly flags that this may not generalize to other speakers — her practical advice for convincing organizers is to study a event's past talk titles and topics, write an outline tailored to that specific audience, and substitute another video or livestream if no prior conference talk footage exists. ## 🗒️ Notes - Agency-to-product pivot is dated precisely: Andy leaves client work January 2016; first plugin ships March 2016 and gets fast sales. The multi-year gap between "talking about selling products" and "actually building them" is attributed to client work crowding out internal projects, not to any product or market uncertainty. - Remote-team claim has a specific inflection point: worked as a small, founder-run team "for a few years," then sales growth made solo support/dev delivery unrealistic. Keith frames the broader WordPress-space shift to remote-normal as fortunate timing rather than something Barn2 drove. - CEO time-audit specifics: ~18-person team at the time; findings were (a) too much time on disliked tasks, (b) too many low-level tasks that should've been delegated to someone junior. Two named structural fixes: delegate specific low-level tasks, hire a marketing coordinator. - AI/website-building claim is explicitly qualified, not absolute: "only a small proportion of people will actually want to build their own website with AI" — the larger group uses AI-assisted builders on existing platforms like WordPress, which is framed as preserving WordPress's security/plugin-ecosystem advantage. - AI/vibe-coding-as-plugin-substitute is a separate, narrower claim from the website-building one: irresponsible for business-critical plugins, high exposure for "minor features" — i.e., small utility plugins, a category Barn2 says it owns several of. - AI search claim is the interview's single most load-bearing, falsifiable statement: two years of declining new sales, attributed to organic-to-AI-search migration, with a named mechanism (AI-recommended traffic converts worse than direct-organic traffic because trust signals on Barn2's own site don't transfer into an AI answer). - Response to the sales decline is threefold and explicitly not confined to marketing-channel swaps: (1) add AI-recommendation optimization alongside existing SEO/content, (2) build non-search channels (cross-promotion, partnerships), (3) diversify products (Shopify apps, documentlibrary.pro) — and Keith states plainly that channel-only marketing changes "will not recover our lost sales" on their own. - Shopify-app distribution mechanism is structurally different from WordPress-plugin distribution: "the vast majority of installs come from the official app store," which Keith frames as protection from organic-SEO-dependent threats — an explicit reason for the platform diversification, not just an unrelated growth bet. - International-sales section names two concrete compliance surfaces: per-country sales tax (solved via a Barn2-built EDD plugin) and translation/localization readiness (special characters, right-to-left support) — despite which the company's own site stays English-only. - Public-speaking escalation path is stepwise and named in order: joint talk → filmed practice posted on X → ten-minute solo lightning talk → full-length solo talk at a smaller event. The "don't overthink mannerisms" conclusion is drawn from a specific observed contradiction: she studied speaker mannerisms on YouTube, then found she used the same mannerisms unconsciously once she actually filmed herself. ## 📚 References - [Barn2](https://barn2.com/) — company site, named as the interview subject's employer. - [documentlibrary.pro](https://documentlibrary.pro) — cited in-article as Barn2's cloud/SaaS product. - No other outbound sources, studies, or data citations appear in the article; every claim about sales trends, team size, or timelines is Keith's own first-hand account with no linked external evidence. ## 🗣️ Paraphrase Katie Keith describes Barn2's story as a series of deliberate trade-offs rather than a single strategy: trading client-services flexibility for product risk in 2016, trading office cohesion for remote-hire flexibility as the company grew past what she and her husband could run alone, and trading her own hands-on control over low-level tasks for delegation once a personal time audit showed her doing work a junior hire should own. The throughline connecting the AI-search, Shopify, and SaaS sections is that Barn2's growth model — built for a decade on ranking well in Google organic search and converting that traffic on its own site — has been structurally undermined for two years, and the company's response is to stop depending on any single channel: add AI-recommendation visibility next to SEO, build products that distribute outside search entirely (an app store, direct SaaS demand), and lean on relationship-based channels like cross-promotion and partnerships that don't route through a search engine at all. The public-speaking section closes the interview on a different register — personal advice rather than business strategy — but shares the same underlying pattern as the rest: start with the smallest viable version (a joint talk, a ten-minute lightning talk) before committing to the full version, and trust direct evidence over secondhand technique (discovering her own mannerisms were fine once she actually recorded herself, rather than continuing to study others'). ## 📰 Newsletter Barn2 co-founder Katie Keith gave 20i a candid, numbers-adjacent look at what AI-mediated search is actually doing to a decade-old WordPress plugin business: two years of declining new sales, attributed to AI recommendations converting worse than direct Google-organic visits because the trust signals on a company's own site don't survive translation into an AI answer. Her response isn't "write more content" — it's platform diversification (Shopify apps, a SaaS wrapper around an existing plugin) plus non-search channels like partnerships and cross-promotion, because she's explicit that SEO and AI-recommendation work alone won't recover what's been lost. Worth reading alongside anything you're planning for plugin discoverability or product diversification this cycle. She also flags small utility plugins specifically as the WordPress product category most exposed to AI vibe-coding — worth a look if that's your product line. Bonus: a genuinely practical public-speaking on-ramp (joint talk → filmed practice → ten-minute lightning talk → first solo talk) for anyone building toward a WordCamp submission. ## ✅ Action Items #### DevRel - Not applicable to this source — the interview is a founder profile with no developer-advocacy program, talk-tracking, or DX-tooling content. #### OSPO - Not applicable to this source — no licensing, governance, SBOM, or security-disclosure process is discussed; Barn2 is a commercial plugin vendor, not an open-source project maintainer, in this interview. #### Community - Not applicable to this source — the remote-team content describes an internal company team (ClickUp chat, a monthly social call), not a broader open-source contributor or sponsor community. #### WordPress - **Plugins:** Keith's claim that small utility plugins face the highest AI-vibe-coding exposure is a concrete, checkable signal for the wp.org Plugins team — worth cross-referencing against actual submission/update volume in low-complexity plugin categories over the next few release cycles. - **Marketing:** Barn2's stated pivot toward optimizing for AI-recommendation visibility (alongside, not instead of, SEO) is a narrative WordPress's own marketing/evangelism efforts may want to track, since the same organic-to-AI-search shift affects wordpress.org's own plugin-discovery funnel, not just individual vendors. - All other subteams (Core, Core Performance, Core AI, Core Program Team, Mobile, Playground, Community, Training, Support, Design, Documentation, Photos, TV, Polyglots, Accessibility, Meta, CLI, Hosting, Tide, Openverse, Themes) — not applicable; the interview never touches core development, infrastructure, or governance. #### The WPOCC - Not applicable to this source — no fellowship, funding, or governance content; Barn2 is discussed strictly as a commercial vendor. #### FAIR - Not applicable to this source — no federation, provenance, or plugin-governance-artifact content appears. ## 🧪 Hypotheses - H1: The shift from organic search to AI-mediated search reduces conversion rate specifically (not just traffic volume) for commercial WordPress plugin sites, because trust signals present on-site (CTAs, newsletter signups, professionalism cues) don't transfer into AI-generated answers. - H2: WordPress plugins classified as "small utility" (single, narrow feature; low implementation complexity) show a measurably higher rate of AI/vibe-coded substitution than plugins solving business-critical or complex functionality. - H3: Revenue channels that don't route through a search engine (app-store distribution, direct SaaS signup, partnership/cross-promotion referrals) are more resistant to AI-search disruption than organic-search-dependent channels, over the same time window. - H4: A founder-level time audit that surfaces undelegated low-level work reliably produces measurable delegation and hiring changes within the following year, independent of company size at the time of the audit. - H5: Practicing a talk by recording and reviewing yourself produces speaking mannerisms indistinguishable from ones learned by deliberately studying other speakers' technique. ### 📊 Methodology - H1: A/B or cohort-compare conversion rate (visit → purchase) for traffic Barn2 can tag as AI-referred (e.g., via referrer strings from known AI assistants, where available) against organic-Google-referred traffic over a matched time window; control for product line and price point. - H2: Compare wp.org plugin directory submission/update velocity and new-plugin counts, segmented by an independent complexity rubric (LOC, feature count, dependency count) for "small utility" vs. business-critical categories, across quarters before/after mainstream vibe-coding tool adoption. - H3: Track Barn2's (or a comparable vendor's) revenue by channel over the same two-year window Keith cites, comparing year-over-year growth in app-store, direct-SaaS, and partnership-referred revenue against organic-search-referred plugin revenue. - H4: Survey or interview a sample of solo/small-team SaaS or plugin-business founders who've run a personal time audit; measure whether delegation/hiring changes followed within 12 months, and whether team size at audit time predicts the outcome. - H5: Blind-rate video clips of Keith's early filmed practice sessions against her later live-talk footage on a standard presentation-mannerism rubric, scored by reviewers unaware of which clip is "practice" vs. "after studying other speakers." ### ❓ Questions - What share of Barn2's traffic is actually AI-referred versus organic-Google, and has that share itself been rising over the same two years sales have been declining — or could a different factor (competition, market saturation, pricing) explain part of the decline? - Does the "AI recommends us but converts worse" mechanism hold across all of Barn2's product lines, or is it concentrated in specific plugin categories (e.g., the same small-utility plugins Keith separately flags as most vibe-coding-exposed)? - Is Shopify app-store insulation from organic-SEO threats durable, or does it just relocate the same discoverability risk to a different single-channel dependency (Shopify's own search/ranking algorithm)? - How does Barn2 define "business-critical" versus "minor feature" when judging which of its own plugins are vibe-coding-exposed, and would that boundary hold up under a customer's actual usage rather than the vendor's own categorization? ### 🔮 Future Research - No published data accompanies either headline claim (the two-year sales decline, or small-utility-plugin exposure) — both are stated as Keith's belief/first-hand account, not measured against Barn2's own analytics or a competitor benchmark. - No comparison to other WordPress plugin vendors experiencing (or not experiencing) the same organic-to-AI-search decline over the same window, which would help separate a company-specific problem from an industry-wide one. - No follow-up on whether the AI-recommendation-optimization effort Keith describes as a planned response has actually shipped or shown results, since this interview captures the plan at the point of stating it, not an outcome. - No treatment of whether Barn2's own SEO/content practices (site structure, backlink profile, content freshness) changed independently over the same two years in ways that could confound the AI-search attribution. ### 💡 Implications - **Practice.** Plugin vendors relying on Google-organic as a primary acquisition channel have a concrete, named peer example of that channel degrading over a multi-year window — worth auditing referral-source trends now rather than after a similar decline compounds. - **Policy/governance.** None named directly in the source; the wp.org Plugins-team connection under Action Items is inference from the vibe-coding claim, not something Keith calls for herself. - **Security.** Keith's own distinction — irresponsible to vibe-code a business-critical plugin, acceptable for minor features — is a self-imposed line with no external verification mechanism named; nothing in the interview describes how a buyer would know which category a given AI-assisted plugin falls into before installing it. - **Education.** The public-speaking escalation path (joint talk → filmed self-review → short solo talk → full solo talk) is a reusable, low-cost on-ramp independent of the WordPress-specific content around it. - **Funding models.** Barn2's platform diversification (Shopify, SaaS) funded from existing plugin revenue is a self-funded response to channel risk, not an externally funded or grant-based one — relevant as a contrast case for vendors considering the same move without Barn2's decade of existing plugin revenue to draw on. ## 🧠 Learning Styles Assessment - **Reading level estimate:** Flesch-Kincaid Grade Level ≈ 9. Basis: mostly compound rather than complex sentences, first-person conversational register, short paragraphs (often one to three sentences), and low use of specialized jargon outside a handful of named terms (vibe coding, SaaS, EDD). #### 👀 Visual learners - **Present:** a header photo, a team photo, and one embedded video clip; no diagrams, charts, or labeled screenshots anywhere in the piece. - **Improve:** a simple before/after diagram of Barn2's revenue mix (organic-search-dependent vs. diversified channels, pre- and post-2024) would make the sales-decline claim far more concrete than prose alone. #### 🔊 Aural learners - **Present:** one embedded video clip in the public-speaking section; otherwise text-only. - **Improve:** an audio version or podcast cut of the interview would suit a reader-turned-listener audience, especially for the more narrative sections (agency pivot, time audit). #### 🗣️ Verbal learners - **Present:** Keith defines "vibe coding" implicitly through examples (her own side project, business-critical vs. minor-feature plugins) rather than a formal definition. - **Improve:** an explicit glossary line for "vibe coding" and "AI search" on first use would help readers unfamiliar with either term follow the AI sections without inferring meaning from context. #### 🧩 Kinesthetic learners - **Present:** the public-speaking section functions as a literal step-by-step path a reader could follow directly. - **Improve:** the marketing-diversification section (AI-recommendation optimization, cross-promotion, partnerships) stays abstract — a checklist of concrete first actions would make it actionable rather than descriptive. #### 🧠 Logical learners - **Present:** the interview is structured as a clean Q&A with one topic per question, which itself creates a logical, scannable structure. - **Improve:** a table mapping each named threat (AI website builders, vibe-coded plugins, AI search) to Barn2's specific response to each would make the cause-and-response logic explicit rather than distributed across several answers. #### 🧑‍🤝‍🧑 Social learners - **Present:** none — this is a one-on-one interview with no comments, discussion prompts, or community engagement shown. - **Improve:** a closing question inviting other WordPress plugin vendors to share their own organic-to-AI-search experience would extend the piece into a comparable-data-gathering exercise. #### 🧘‍♀️ Solitary learners - **Present:** the piece works well as a single, linear read; each Q&A section is self-contained. - **Improve:** a short "if you run a plugin business, ask yourself" reflection prompt after the AI-search section would help a solitary reader self-apply the diagnosis to their own numbers. #### 🌿 Naturalistic learners - **Present:** none — no ecological, sustainability, or systems-in-nature framing appears anywhere in the article. - **Improve:** not a natural fit for this content; no changes recommended. ## 📊 Accessibility Assessment #### Screen reader semantics - **Present:** the Q&A structure uses bold question headers, giving clear topic segmentation that would translate well to proper heading markup. - **Improve:** every `<figure><img>` in the captured HTML carries an empty `alt=""` attribute — the header image, the team photo, and the pull-quote figure all lack descriptive alt text (WCAG 2.2 SC 1.1.1 Non-text Content). #### Keyboard navigation - **Not observed** — no interactive elements beyond standard links and an embedded video are present in the captured content; full keyboard-trap or focus-order testing would require live-page review, not the captured HTML alone. #### Contrast and readability - **Not observed** — text/background contrast isn't verifiable from captured HTML alone; flag for direct visual review before treating as compliant (WCAG 2.2 SC 1.4.3). #### Text resizing and reflow - **Not observed** — no fixed-width or overflow-prone elements appear in the captured content; would need live-page testing at 200% zoom to confirm (SC 1.4.4, SC 1.4.10). #### Media and captions - **Not observed for the embedded video** — the captured HTML shows only an `<embed>` reference to a YouTube video with no caption-track information; whether the video itself has accurate captions can't be confirmed from this capture (WCAG 2.2 SC 1.2.2). #### Forms and interactions - **Not applicable** — no forms or interactive inputs appear anywhere in the article. #### Standards reference - Confirmed gap: WCAG 2.2 SC 1.1.1 (all three images ship with empty alt text). Unconfirmed, needs live review: SC 1.2.2, 1.4.3, 1.4.4, 1.4.10, 2.1.1, 2.4.3. ## 🌍 DEIB Assessment #### Cultural breadth - **Present:** Keith's own geography spans southwest England and Mallorca, and she frames Barn2's remote model as enabling international hiring "wherever they are in the world"; the interview also covers Barn2's international sales localization work. - **Improve:** no specific countries, languages, or team-member backgrounds are named beyond Keith's own location — a published version could name the actual spread of the ~18-person team without disclosing individuals' identities. #### Inclusive language - **Present:** consistent first-person, plain register throughout; no idiomatic phrasing likely to translate poorly. - **Improve:** none needed; the interview's language is already direct. #### Neurodiversity support - **Present:** the Q&A format itself chunks content into predictable, single-topic units, which aids scanning and reduces cognitive load compared to unstructured narrative. - **Improve:** several answers run to 6–8 short paragraphs without internal subheadings (notably the AI/vibe-coding and AI-search answers) — breaking these into labeled sub-points would help readers who rely on structure over continuous prose. #### Representation - **Present:** Keith explicitly names her own position as "a woman in tech who owns a reasonably well established company" when discussing how she convinced event organizers to book her — a direct, self-reported representation data point rather than an inferred one. - **Improve:** she also states this "might be harder for other people" without elaborating on whom or why; a follow-up interview naming which groups face more friction getting booked would extend this into something actionable rather than a passing acknowledgment. #### Psychological safety - **Present:** Keith names her own past anxiety about public speaking directly ("If you're not confident with public speaking...") without minimizing it, and frames starting small as a legitimate strategy rather than a compromise. - **Improve:** none needed for this content; no sensitive or high-stakes material requiring a content warning appears. ## ✅ Logical Validity #### Strong reasoning - **Causal mechanism on AI-search conversion (organic-vs-AI-referral section):** the claim that AI-referred visitors convert worse than direct-organic visitors because trust signals (CTAs, newsletter signups, professionalism cues) don't carry into an AI-generated answer is a specific, testable mechanism, not a bare assertion — it names exactly what's lost in the AI-mediated path. - **Structural reasoning on Shopify-app insulation:** tying app-store-majority installs to reduced organic-SEO dependence is a clear, mechanistically sound argument — it identifies precisely why one distribution channel is less exposed to the same risk as another, rather than just asserting diversification is generally good. #### Fallacies - **Hasty dismissal of a counterargument.** Location: "A lot of people try to deny this by pointing out that it's hard to maintain those sites once AI has built them. I think that argument is irrelevant, because the AI can easily make changes, so the website owner never has to learn how to update their site." Why it's weak: this addresses only the *labor* cost of maintenance, not the counterargument's likely deeper concern — the quality, security, and correctness of AI-driven changes over time to a site the owner doesn't understand. Calling the argument "irrelevant" closes the question rather than engaging its strongest form. Fix: acknowledging that "easy to change" and "safe to change" are different claims would strengthen the point rather than weaken it. - **Unsupported single-cause attribution.** Location: "Our new sales have been declining steadily for two years, and I believe this is the reason." Why it's weak: Keith flags this herself as belief, not measured fact, but the rest of the answer proceeds as if the causal link were established — no alternative explanation (pricing, competition, market saturation, product-market shifts) is considered or ruled out. Fix: the claim would be stronger paired with even one comparison point, such as a competitor's traffic trend or Barn2's own referrer-source breakdown over the same window. #### Tone of voice - First-person, direct, and unhedged on business decisions ("we need to," "I believe," "I think"); notably more hedged and self-aware in the public-speaking section ("I recognize that it might be harder for other people"); no defensive or evasive phrasing anywhere in the piece. ## 🛡️ DARVO Analysis - **Deny:** Not applicable to this source — no accusation exists against Keith or Barn2 for the DARVO framework to apply to. Where she addresses a counterargument (the AI-site-maintenance objection), she is rebutting a technical claim about AI capability, not denying a wrongdoing allegation. - **Attack:** Not applicable to this source — no critic or accuser is named or attacked; "a lot of people" who raise the maintenance objection are referenced generically and their argument is addressed, not their credibility. - **Reverse victim/offender:** Not applicable to this source — no victim/offender framing appears anywhere in the interview; Keith describes Barn2 as affected by a market shift (AI search), not by any party's misconduct, so there is no accused-to-victim reversal to analyze. - DARVO doesn't fit this source because nothing in it involves an accusation of harm against Keith or Barn2. The only defensive-shaped rhetoric present is the counterargument dismissal analyzed above under Logical Validity, which is a reasoning weakness, not a DARVO move. ## 🎯 Conclusion Barn2's story, as Keith tells it, turns on one dated, self-reported inflection point — two years of declining new-plugin sales she attributes to the shift from Google-organic to AI-mediated search — and the company's response is structural rather than cosmetic: diversify distribution (Shopify's app store, a direct SaaS product) and marketing channels (partnerships, cross-promotion, AI-recommendation optimization) rather than trying to out-SEO a channel she says won't fully recover. The second AI threat she names, small utility plugins losing ground to vibe-coded substitutes, is a distinct mechanism from the search-conversion problem and worth tracking separately. Neither claim comes with external data in this interview; a useful next step for a reader running a similar plugin business is to check their own referrer-source and conversion-rate trends over the same two-year window before assuming the same mechanism applies to them. ## 🔄 Refactor Appendix — Interview with Barn2 co-founder Katie Keith - 20i® > Density: 2 Literature atoms (H4), 1 Permanent candidate (H4). Both atoms verified byte-exact > against html_content. Konik claim-header rules: present tense, declarative, ≤10 words, > searchable, unique. Atomicity: one declarative sentence per atom. ### 📚 Literature Atoms (H4) #### [H1] AI search has cut Barn2's new-plugin sales for two years > [!quote] > Our new sales have been declining steadily for two years, and I believe this is the reason. **Context & use**: - Directly follows: "The move from organic search to AI search has been disastrous for Barn2." — this sentence names the specific, dated, falsifiable effect Keith attributes to that shift. - Load-bearing for the whole back half of the interview: Barn2's Shopify-app and SaaS diversification is described elsewhere in the source as a direct response to this same sales decline. **Tags**: #xref #addendum **Highlight**: verified verbatim against `html_content`, doc 01m1pq04nstdt5beg946gzthm3 <!-- atomicity: 1 idea ✓ --> #### [H2] Small utility plugins face the biggest vibe-coding threat > [!quote] > As a result, small utility plugins are under the biggest threat. **Context & use**: - Follows Keith's distinction between vibe-coding a "business-critical plugin" (which she calls irresponsible) and vibe-coding "minor features" (which she says nothing stops people from doing); this sentence names which plugin category absorbs that risk. - Keith immediately adds Barn2 itself owns "several of these," making the claim self-implicating rather than a generic industry warning. **Tags**: #xref **Highlight**: verified verbatim against `html_content`, doc 01m1pq04nstdt5beg946gzthm3 <!-- atomicity: 1 idea ✓ --> ### 🧠 Permanent Candidates (H4) #### [P1] AI threatens plugin businesses through two separate mechanisms **Claim** (Courtney's voice): When I hear "AI is a threat to WordPress plugin businesses," I should ask which of two distinct mechanisms is meant — AI search cutting conversion on existing demand, or AI code generation substituting for the product itself — because they call for different responses. **Evidence pointer**: [[#H1]], [[#H2]] **Scope & conditions**: Holds for any plugin vendor whose acquisition depends on organic search and whose product line includes low-complexity utility features; the search-conversion mechanism is a discovery-layer problem (fixable with channel diversification), while the vibe-coding mechanism is a product-category problem (fixable only by moving upmarket in complexity or defensibility) — treating them as one problem risks applying the wrong fix. **Links** (≥1 with annotation): - [[Plugin-business economics]] — because extends: both mechanisms narrow the addressable market for a specific class of WordPress product business, from different directions. - [[Channel concentration risk]] — because applies-in: any single-vendor business built on one dominant acquisition channel (here, Google organic) is exposed the same way regardless of industry. <!-- atomicity: 1 idea ✓ --> ### ✏️ Fleeting Queue - [F1] Check whether Barn2's AI-recommendation-optimization effort (mentioned as a planned response, not a completed one) has since shipped or reported results. — https://www.20i.com/blog/interview-with-barn2-co-founder-katie-keith/ #xref - [F2] Cross-reference Keith's "small utility plugins under the biggest threat" claim against actual wp.org plugin-directory submission trends in low-complexity categories. — #xref ### → Things 3 (max 2 actions) - [ ] Pull referrer-source and conversion-rate trends for any plugin/product business I track over the past two years and check whether the AI-search-conversion mechanism Keith describes shows up in the numbers — context #plugin-business-research #things - [ ] Read up on Shopify's own app-store ranking/discovery algorithm before treating "app-store distribution" as inherently safer than organic search — context #verification #things <!-- deep-read:sections end --> 📄 **Full text:** [[Full Document Contents/Articles/Interview With Barn2 Co-Founder Katie Keith - 20i®.md|Interview With Barn2 Co-Founder Katie Keith - 20i® by Katie Keith]] ## Highlights - However, there’s nothing to stop people from using AI to build minor features. As a result, small utility plugins are under the biggest threat. Unfortunately, we have several of these at Barn2! ([View Highlight](https://read.readwise.io/read/01m1sj660v0w11h3y1sf2qefpm)) - Note: H3 -- Supporting context: Keith names her own product category (small utility plugins) as the one most exposed to AI-assisted vibe coding, an admission against her own commercial interest. #atom/literature/h3 - Tags: atom/literature/h3 - The move from organic search to AI search has been disastrous for Barn2. Our new sales have been declining steadily for two years, and I believe this is the reason. ([View Highlight](https://read.readwise.io/read/01m1sj61zahkbfmvqst76mx2sr)) - Note: H1 -- Keith names AI-mediated search, not general market softness, as the cause of two years of declining new-plugin sales at Barn2. #atom/literature/h1 - Tags: atom/literature/h1 - When people arrive straight on our site from Google, they are likely to make a purchase after seeing our professionalism, trust factors, calls to action, newsletter signups and all the other things that don’t appear in AI recommendations. ([View Highlight](https://read.readwise.io/read/01m1sj63tmxa1n5q3416rbm20p)) - Note: H2 -- The mechanism behind H1: an AI answer strips out the trust signals that convert a Google visitor into a buyer, so an AI citation is worth less than an organic click even when volume looks similar. #atom/literature/h2 - Tags: atom/literature/h2 ## New Highlights added 2026-09-10 at 10:41 AM - As a result, small utility plugins are under the biggest threat. ([View Highlight](https://read.readwise.io/read/01m25w08g2xcyd197rgqa30jtn)) - Note: The threat category named. Compare WP Product Talk (same batch), which independently reaches the same segment without the sales figure. - Tags: wordpress-plugins, xref - Our new sales have been declining steadily for two years, and I believe this is the reason. ([View Highlight](https://read.readwise.io/read/01m25w056wrdzbhss5kffqnjyb)) - Note: Keith's own first-hand attribution — no external data, competitor comparison, or Barn2 analytics cited. - Tags: xref, wordpress-plugins