Safe Doesn't Scale

David Walsh

"What's the ROI?" Those three words kill more creative marketing ideas than bad execution ever will. Not here. Safe Doesn't Scale is a weekly podcast for marketing and growth leaders. We’ll be interviewing Heads of Marketing, Unicorn Founders, and Revenue Leaders at B2B companies to prove that the riskiest marketing campaigns drive the biggest returns. While brands are burning $500K on LinkedIn ads that are generating zero demos, there’s someone out there who closed a $2M deal they sourced from a meme. Host David Walsh, Founder of Limelight, breaks down real examples from brands spending less and converting more by leaning into creator-led growth, unconventional distribution, and campaigns that make traditional marketers panic. You’ll learn: How growth leaders sell “unsafe” ideas to the C-suite How to attribute sales pipeline to content, creators, and social signals Why the campaigns that feel uncomfortable often drive the most revenue No e-book downloads. No buzzwords. This show is for marketers with a chip on their shoulder who are tired of playing it safe. We celebrate the campaigns that make legal sweat and sales teams crush quota. Because marketers who don't take risks won't exist in 2027.

  1. 27 Aug

    AI Became Public Enemy Number One in B2B Social (with Meryoli Arias, ex-Apollo and Chili Piper) | Ep. 22

    AI Made Content Cheap. It Made Trust Expensive. ㅤ Every B2B company now has the ability to publish more than it ever has, and buyers are reading less of it than they ever have. LinkedIn shipped a button that lets people flag posts as AI slop, which tells you where the feed ended up. The companies still getting attention are doing something that can't be generated. ㅤ David Walsh, founder of Limelight, talks with Meryoli Arias, who built social and community programs at Chili Piper and Apollo before going independent. They cover what to publish when everyone can publish, how to run employee advocacy without mandating it, and how to prove social is working when attribution won't cooperate. ㅤ Meryoli has run the version of this that worked at scale: an employee advocacy program at Chili Piper five years before it was common, a nomination campaign that pulled in millions of impressions in a month, and a stretch at Chili Piper where more than 70% of closed deals said they first heard about the company through social. ㅤ Guest Bio Meryoli Arias is a B2B social media and community marketing consultant who now works with earlier-stage companies setting up their first real social strategy. She spent six years inside B2B marketing teams, including Chili Piper, where she ran social and helped build the employee advocacy program the company became known for, and Apollo, where she led social through the period right before the company hit unicorn status. Her work centers on what she calls community marketing: activating customers, partners, and employees rather than broadcasting from a company page. She still spends most of her day on LinkedIn, and she is one of the few marketers who will say out loud that she likes it there. ㅤ What We Cover Why social went from nice to have to table stakes: Products, websites, and copy can be replicated quickly now. Brand and the people behind it can't. Meryoli traces how social moved from afterthought to competitive advantage to baseline requirement.What she saw working at Apollo: She joined when the company started investing in brand, right before unicorn status. The product already had organic advocates, so the team started conversations with those people instead of talking over them.AI as public enemy number one: The mistake she sees most is leadership believing social strategy can be replaced by AI. Production can be automated. The judgment about what to publish can't.The LinkedIn AI slop button: People stopped reading posts because they assumed AI wrote them. LinkedIn responded to that behavior with a reporting mechanism, and Meryoli says the feed has shifted noticeably in the three weeks since.Why inauthentic content gets discounted: She draws a line between AI content and paid endorsements. Once you know something was paid for, you pull back. The same reflex now applies to anything that reads as machine-written.Employee advocacy without the mandate: Handing employees copy to paste, or requiring reposts, produces nothing. She starts by showing individuals what a personal brand does for their own career, then works with the people who opt in.Building criteria instead of a content calendar: Her term for the internal standard someone applies to a draft. AI can produce the first version. Criteria is what tells you how to make it better, and it only comes from experience.Who a founder should actually hire first: A founder asked her whether his first marketing hire should be a social specialist. Her answer was no. Start with a generalist who understands a bit of everything, and start the social foundation early regardless.Community marketing versus building a community: Two different jobs. Building a community means owning the space and keeping it alive. Community marketing means joining conversations that already happen in your category, on Reddit or inside groups your buyers are in.Choosing channels beyond LinkedIn: The first question is where your ICP spends time. Engineers lean toward Reddit and YouTube. Designers go somewhere more visual. She has seen results on TikTok, Instagram, Reddit, and the recent revival of X.Measuring social when attribution breaks: Nobody sees a post and buys. Her preferred signal is the sales team reporting that prospects say they heard about the company on social, plus the answer to the onboarding question about where they first heard of you.Campaigns worth stealing: The Chili Piper top 32 marketers nomination campaign, where nominees rallied their own followers to vote, and the employee takeover format, where an employee wrote a post for the company page and, on launches, the engineer who built the product explained why. ㅤ Resources Mentioned Apollo: Where Meryoli ran social during the run up to unicorn status, including an engineering newsletter written by the engineers themselves.Chili Piper: The employee advocacy program, the employee takeover format, and the top 32 marketers campaign all came out of her time there.Exit Five: The B2B marketing community Meryoli recommends, whose event David hears is sold out and whose newsletter he already reads.Dave Gerhardt: Founder of Exit Five, and the writer of one of the only newsletters David subscribes to.Gong: David describes teams running agents across sales call transcripts to count how often prospects mention a social post, an employee, or a creator.Reddit: Named as a place where category conversations already happen, and a strong channel for engineering audiences.ChatGPT: Meryoli's example of a first draft that experience lets you improve, rather than a finished post.Notion: Home of David's to-do list, where joining Exit Five has been stuck in progress. ㅤ Safe Doesn't Scale is hosted by David Walsh, founder of Limelight. New episodes drop weekly.

    AI Became Public Enemy Number One in B2B Social (with Meryoli Arias, ex-Apollo and Chili Piper) | Ep. 22
  2. 6 Aug

    The Best Influencers Aren't Influencers (with Max Nimmo from CreatorWorks) | Ep. 21

    The Best Influencers Aren't Influencers. B2B teams will happily pour money into paid social for a year before asking it to prove itself. Give a creator program three months and one post, and the same team wants pipeline by Friday. That double standard is quietly killing a channel before most companies have run it properly once. ㅤ David Walsh, founder of Limelight, sat down with Max Nimmo, who started CreatorWorks a year ago and now runs creator programs for B2B SaaS and AI companies. They get into what a creator budget actually has to be, why organic content on its own stopped producing results, and how to turn creator posts into thought leader ads without blowing up your rates. You'll leave with a number to benchmark against and a clear view of where creator content sits in a go-to-market motion. ㅤ Max came into B2B from gaming, where he ran campaigns with YouTubers and Twitch streamers for brands like Activision Blizzard, then moved to the performance side building marketing mix models for influencer spend. He brings both halves to this conversation: the creative relationship work and the measurement discipline that most influencer programs skip. ㅤGuest BioMax Nimmo is the founder of CreatorWorks, a creator-led growth agency for B2B SaaS and AI companies, which he started in London a year before this recording. He spent his early career in gaming influencer marketing, working with YouTubers and Twitch streamers on campaigns for brands including Activision Blizzard and Call of Duty, before moving to the performance marketing side and building marketing mix models that measured influencer spend against paid media. At CreatorWorks he runs a small team supported by an automated middle layer, managing pods of five to 10 clients across LinkedIn creator programs, thought leader ads, and intent tracking. Before he founded the company, he hadn't logged into LinkedIn in about five years. He describes the version of himself that stumbled into the industry as someone playing video games 16 hours a day at university with no idea what to do next. ㅤ What We CoverWhat creator-led growth means: Max separates the sponsored post everyone recognizes from the wider motion he sells. Organic creator content sits at the foundation, then thought leader ads, engagement tracking, and signal capture turn it into a loop rather than a one-off campaign.Why organic alone stopped working: Partnerships that rely purely on organic reach aren't producing the results brands expect anymore. Max is direct that impressions and CPM are the wrong scorecard, and that consistent results now require going beyond organic.From Call of Duty to LinkedIn: Gaming adopted influencer marketing early and understood creator distribution before anyone else. Max traces the line from managing campaigns with Twitch streamers to building marketing mix models to landing in B2B, where product creation is commoditized and distribution is what's left.The 10/80/10 agency model: The first 10% of any campaign is human strategy. The middle 80% runs on agents, workflows, Airtable, and Claude Code. The final 10% is a human check before anything reaches a client or goes live, which is how a very small team supports a pod of five to 10 clients.The real budget floor: Max reverse engineers from the number of creators, pieces of content, and data points needed to see results in three months. That lands at 10 to 15K per month as an absolute minimum for LinkedIn, before ad spend or usage rights.The measurement double standard: Nobody expects one Meta ad or one event to generate a flood of leads, but a single creator post gets held to exactly that. David adds that sales cycles often run six to nine months, longer than most companies have even been running influencer.The B2C playbook mostly transfers: Roughly 80% of what Max ran for D2C brands is what he now runs in B2B. The differences sit at the back end: tying content to a CRM, routing intent to SDRs. He calls the wider "you can't apply consumer playbooks here" argument rubbish.The best influencers aren't influencers: Founders, analysts, investors, and people doing the work day to day carry the most influence, and none of them think of themselves as creators. That changes the working relationship, because they need real input on hooks and formats rather than a brief and a wave goodbye.Reading the algorithm in real time: Max uses Favikon to find top-performing posts and creators, then collects CSVs from creators because LinkedIn's own reporting doesn't give enough. He describes a creator with over 100,000 followers whose lead magnet posts suddenly couldn't clear 200 impressions.Big creators versus small creators: David's experience is that brands reach for the largest names by default and get worse outcomes. Max pushes back on the binary: different creator sizes serve different jobs and should be measured on different things, but budget forces a choice.Usage rights get negotiated on day one: Thought leader ads run about 3X the click-through rate of regular brand creatives, so repurposing rights are agreed up front with an execution clause. Wait until a post performs and the creator's rate goes up immediately.Authenticity is overrated as a first principle: Max says it's very important but not the most important thing, and questions how well any of us actually detect it. The fix is positioning: not "I use this and love it" but "I found this recently and here's why it matters to you."Creators as a signal engine: A Clay strategy needs quality data to scrape from the outset. Tracking your competitor's posts or your founder's five posts a week produces a trickle. A creator program produces thousands of data points to feed a CRM, outbound sequences, or ads.The AI visibility hot take: Companies run five tools to measure AI visibility because nobody has an accurate answer yet. What's known is that third-party sources like LinkedIn, Reddit, and YouTube get cited most, and everyone else is guessing with confidence. ㅤ Resources MentionedFavikon: the tool Max returns to for finding top-performing LinkedIn posts and creators.Clay: named as the destination for intent signals, which is why the quality of the data you scrape matters at the outset.Airtable: part of the CreatorWorks stack for building internal dashboards and campaign workflows.Claude Code: used alongside Airtable to automate the middle 80% of campaign execution.Y Combinator: its request for AI-enabled service businesses is the thesis CreatorWorks was built on.20VC: David references Harry Stebbings' observation about how many founders were once gamers. ㅤ Safe Doesn't Scale is hosted by David Walsh, founder of Limelight. New episodes drop weekly.

    The Best Influencers Aren't Influencers (with Max Nimmo from CreatorWorks) | Ep. 21
  3. 16 July

    The $50,000 Floor Nobody Tells You About Before You Try Influencer (with Stephen Titus from Faved) | Ep. 20

    Most B2B teams have already run their influencer test. Four or five creators sitting dead center in the category, twenty or thirty thousand dollars, a handful of posts, and a quiet decision that the channel doesn't work. The test was never big enough to prove anything either way. ㅤ David Walsh, founder of Limelight, sits down with the co-founder of a competing marketplace to argue about where creator budgets actually break. You'll walk away with a spending floor, a reason to buy creators outside your category, and a straight answer on what AEO changes about influencer content. ㅤ Stephen brings four years of running flat-fee creator partnerships at scale, a bootstrapped P&L that forces him to care what brands actually renew on, and a blunt estimate of how little of B2B advertising currently reaches creators at all. ㅤ Guest BioStephen Titus is co-founder and CEO of Faved, a marketplace connecting brands with creators for flat-fee sponsorships. He runs it with a team of eight, more than 30,000 creators, and hundreds of brands, and has deliberately kept the company bootstrapped after raising venture capital for earlier ventures. He and his co-founder are engineers who first met the world of influencer marketing by accident: they built a productivity smartwatch straight out of university, ran a Kickstarter, and discovered that niche communities could drive volume that paid media couldn't. Multiple failed startups followed before Faved. He now works with consumer brands in supplements and cosmetics alongside a growing roster of AI and software companies. ㅤ What We CoverTwo founders, two watch companies: Both David and Stephen started watch brands and backed into influencer marketing the same way.Why flat fee, not commission: Stephen's case for making creator deals more transactional and merit-based.The sin of raising VC too early: Investors became the customer, and the actual customer got ignored.Bootstrapped as a forcing function: Revenue as the funding model ties the platform's goals to the brand's goals.The real competitor is an agency: Not platforms. Mostly no-name agencies and headcount.Alignment before spend: Awareness, clicks, qualified interest, or sales. Pick one before the budget moves.The $50,000 floor: Below that, on one platform, across 20 to 30 creators, you learn nothing.What the first 50K actually buys: One or two creators worth doubling down on with the next 50K.Go non-obvious on category: A CRM should be buying performance coaching and freelancer creators, not just sales creators.AEO, SEO, GEO: Stephen's view that it's all the same job with a new front door.Trust beats topical match in LLM answers: Why a running creator mentioning HubSpot can outrank a sales creator.Consistency over spikes: Monthly baseline spend, or your competitors take the share of attention.B2B is under 1%: Stephen's estimate of creator spend as a share of B2B advertising budget. ㅤ Resources MentionedFaved: Stephen's flat-fee creator sponsorship marketplace.Limelight: David's B2B influencer marketplace, discussed as the friendly competitor.Luma AI: Named as one of the AI brands running creator partnerships.Notion: Cited as a brand reaching prosumer audiences through creators.Skillshare: Example of an advertiser running a known playbook that just needs tooling to scale.Brilliant: Named alongside Skillshare as a scaled creator advertiser.HubSpot: Used in the example of trust outranking topical match in ChatGPT answers.Kickstarter: Where Stephen's first company discovered the power of niche communities. ㅤ Safe Doesn't Scale is hosted by David Walsh, founder of Limelight. New episodes drop weekly.

    The $50,000 Floor Nobody Tells You About Before You Try Influencer (with Stephen Titus from Faved) | Ep. 20
  4. 9 July

    68% of Executives Say This One Thing Would Make Them Hit Their Numbers (with Sangram Vajre from GTM Partners) | Ep. 19

    Most B2B companies are running inbound and outbound because they're comfortable, measurable, and feel like a spreadsheet they already know. That's exactly why the motion that creates the biggest deal sizes and longest customer relationships stays chronically underinvested. ㅤ David Walsh, founder of Limelight, sits down with Sangram Vajre, co-founder and CEO of GTM Partners, to get into the real mechanics of go-to-market: how teams lose alignment, why partner-led motion outperforms everything else, and what it actually takes to build content that 175,000 people read weekly. ㅤ Sangram has been in the rooms where these decisions get made. He ran marketing at Pardot through a $2.5 billion Salesforce acquisition, co-founded and scaled the ABM platform Terminus to over $100M in revenue, and now runs the go-to-market research firm he built from zero to ~$10M without raising a single dollar. ㅤ Guest Bio Sangram Vajre is co-founder and CEO of GTM Partners, a data-driven go-to-market analyst and advisory firm he built alongside Bryan Brown, Lindsay Cordell, and Judd Borakove. Before GTM Partners, he co-founded Terminus, one of the companies that created the ABM software category, and before that ran marketing at Pardot through its acquisition into Salesforce. He's the author of three books on B2B go-to-market, including MOVE, a Wall Street Journal and USA Today bestseller co-written with Bryan Brown. He's spoken at UNBOUND (formerly INBOUND) for ten consecutive years and hosts the GTMonday research newsletter, now published as Run on GTM OS, with over 175,000 subscribers. ㅤ What We Cover Why marketing is a silo of silos: The graphic designer, the demand gen manager, the ABM lead, and the social team often have no shared understanding of what they're collectively trying to accomplish.Renaming meetings as a GTM alignment tool: A "marketing attribution meeting" signals the conversation is about justifying marketing's existence — rename it "pipeline velocity" and the whole orientation shifts.How Henry Schuck runs alignment at ZoomInfo: He replaced weekly executive meetings with a daily 9 AM go-to-market session: no decks, no status updates, just the team watching how decisions get made in real time.68% of GTM executives say clarity would make them hit their numbers: When given a list of external pressures, nearly 68% of executives said clarity on their own go-to-market strategy would matter more than fixing competitors, AI, or macroeconomics.The six go-to-market motions and why partner-led wins: Of the six motions GTM Partners tracks, partner-led produces the largest deal sizes and highest business value — and is chronically underfunded because it takes the longest to build.The Salesforce and Bombora playbooks for partner-led growth: Bombora skipped building a sales team entirely, embedded its intent data inside Terminus, Demandbase, and 6sense, and took a revenue share on every deal.Story first, content format second: The CEOs who build the strongest brands are the story — what works for successful creators isn't the posting cadence, it's that they have something real to say.How GTMonday grew to 175,000 subscribers without paid ads: GTM Partners wrote original research every week instead of blog posts, and one day 600 IBM employees subscribed because someone dropped the link in an internal Slack channel.LinkedIn as an LLM and the case for commenting over posting: Comments on high-followership accounts can generate more impressions than standalone posts — for anyone starting out, commentary is a faster path to visibility.The bet on GTM fractionals as the next wave: Sangram's thesis is that operators in their 40s and 50s who feel disrupted by AI are sitting on their most valuable asset: two decades of judgment that AI execution can now amplify. ㅤ Resources Mentioned GTM Partners: Sangram's go-to-market analyst and advisory firm, home of the GTM Operating System framework. Run on GTM OS: GTM Partners' weekly research newsletter with over 175,000 subscribers, formerly published as GTMonday on Substack. ZoomInfo: B2B data and intelligence platform whose CEO Henry Schuck recently changed the company's Nasdaq ticker from ZI to GTM; mentioned as a GTM Partners client and daily alignment case study. HubSpot: B2B CRM and marketing platform mentioned as an example of a company that scaled by making partners central to its revenue motion. Bombora: B2B intent data company that grew without a direct sales team by embedding its data inside ABM platforms and taking a revenue share. Demandbase: ABM platform mentioned as one of the partners that distributed Bombora's intent data. 6sense: Account-based marketing and revenue intelligence platform, also a Bombora distribution partner. PartnerStack: Partner relationship management platform David Walsh considered building before launching Limelight, and a current Limelight integration. impact.com: Partnership and affiliate management platform mentioned alongside PartnerStack as a reference point for the partner-led motion category. Beehiiv: Newsletter platform GTM Partners migrated to from Substack for segmentation and support at scale. Substack: Publishing platform where GTMonday was originally hosted before the move to Beehiiv. ㅤ Safe Doesn't Scale is hosted by David Walsh, founder of Limelight. New episodes drop weekly.

    68% of Executives Say This One Thing Would Make Them Hit Their Numbers (with Sangram Vajre from GTM Partners) | Ep. 19
  5. 2 July

    Your Brand Page Got Deleted. Now What? (with Katie Parkes from Apollo.io) | Ep. 18

    Most B2B brands treat social as a channel they own. Then something happens that makes them realize they don't own it at all. What they actually own is trust, and trust lives with people, not pages. ㅤ David Walsh, founder of Limelight, sits down with Katie Parkes, Director of Social, Community & Customer Marketing at Apollo.io, to talk through what it actually takes to build a social program that compounds. You'll walk away with a concrete framework for customer advocacy at scale, a measurement structure worth stealing, and a sharper view of where to stop spending. ㅤ Katie came into her role at Apollo.io two weeks before the company lost its LinkedIn company page. Rather than treating the loss as a crisis to manage, she used it as a forcing mechanism to rebuild the entire social strategy around individual voices. She's been running that playbook ever since. ㅤ Guest Bio Katie Parkes (Katie Jane Parkes) is the Director of Social, Community & Customer Marketing at Apollo.io, the AI-native go-to-market platform used by 600,000+ companies worldwide. Before Apollo, she was a founding member of Shopify's enterprise marketing team, where she led social media and then video, growing Shopify's social audience by 270% in a single year. She has built creator programs, scaled Slack communities, and run influencer campaigns that track all the way to paid customer revenue. She's based in Canada and will happily talk shop on LinkedIn. ㅤ What We Cover Losing the LinkedIn company page: Apollo's page was removed two weeks into Katie's tenure. She explains what it meant for the brand and how it forced a channel strategy rethink from day one.Diversifying beyond LinkedIn: Katie pushed into Instagram, X, and an owned subreddit. GTM operators were already there; she built awareness from scratch.Individual voices over brand pages: Company pages still matter, but trust gets rebuilt through people. That's where she put her budget first.Finding advocates in a 4-million-user base: Three signals: Slack community activity, UserEvidence survey hand-raisers, and social listening for unprompted mentions.The reciprocal relationship: AI can surface advocates. It can't replace commenting on their content and making the relationship feel real. Most teams skip this at scale.What she's cutting from the budget: Legacy social tools that haven't built genuine AI functionality are getting dropped at renewal. She walks through her build vs. buy framework.How Apollo's marketing team uses AI: A brand agent inside Cognition drafts copy and routes it to a human reviewer. Her team uses Claude Skills for content repurposing and weekly performance pulse checks.Four measurement buckets for social: Growing influence, growing engagement, growing demand, and growing pipeline. Every channel and campaign ladders into one of the four.Tracking influencers beyond impressions: Unique links, session data, and beta signups told her which creators drove demand, not just reach. Those are the ones she renewed. YouTube creators run through PartnerStack and tie directly to paid revenue.The three-bench employee advocacy system: Exec content via a contracted writer. A second bench of employees with audiences or product expertise. A third bench tied to talent acquisition and employer brand. All three activate for major moments.YouTube as a demand channel: She's seen it drive 30% of PLG signups. Apollo's channel has underperformed. A dedicated hire and a clearer strategy are coming in H2, with LLM indexing adding a new reason to take it seriously. ㅤ Resources Mentioned Apollo.io: the AI-native B2B go-to-market platform where Katie leads social, community, and customer marketing. UserEvidence: the customer-evidence and advocacy platform Apollo uses to identify hand-raisers for co-creation and creator programs. PartnerStack: the partner and affiliate management platform Apollo uses to track influencer-driven signups and paid revenue, particularly on YouTube. Cognition: the AI-powered campaign management platform behind Apollo's brand agent and cross-team content calendars. (Could not independently verify the vendor's full name or website; confirm with Katie before publishing.) Claude Skills (Anthropic Agent Skills): Anthropic's repeatable AI workflow feature Katie's team uses for content repurposing and performance reporting. Limelight: David Walsh's B2B influencer marketplace, referenced in the context of launching AI employees for influencer, social listening, employee social, and customer advocacy. ㅤ Safe Doesn't Scale is hosted by David Walsh, founder of Limelight. New episodes drop weekly.

    Your Brand Page Got Deleted. Now What? (with Katie Parkes from Apollo.io) | Ep. 18
  6. 25 June

    Everyone's Obsessed With the Wrong Thing in AI Marketing (with Megan Bowen from Refine Labs) | Ep. 17

    Every marketing team is being told that if they haven't shipped an AI agent this quarter, they're behind. The irony is that the companies winning in the AI era are the ones doing the unglamorous work: clear positioning, real brand investment, messaging that actually says something. The tools changed. What gets you discovered didn't. ㅤ On this episode of Safe Doesn't Scale, David Walsh of Limelight sits down with Megan Bowen, CEO of Refine Labs, to talk about what's actually changed for B2B buyers now that ChatGPT and Claude are part of how they research solutions. They cover positioning, brand measurement, and where AI belongs in a go-to-market strategy. You'll leave with a sharper sense of what to fix before you automate anything. ㅤ Megan spent her career on the operating side of B2B growth before taking over Refine Labs, and she brings receipts. She walks through a controlled LinkedIn experiment that produced 60% more qualified pipeline, why most Google Ads budgets quietly leak money, and the hard lesson Refine Labs learned when it leaned too heavily on a single founder brand. ㅤ Guest BioMegan Bowen is the CEO of Refine Labs, the B2B marketing agency she helped scale before stepping into the top role. Refine Labs launched in 2019 and has worked with more than 300 B2B SaaS and tech companies, driving qualified pipeline through demand strategy, paid media, and content production. Her background runs through customer success and operations leadership at companies like Grubhub and Managed by Q, which is why she treats marketing as a business function rather than a list of tactics. She's become one of the clearer voices arguing that AI raises the value of marketing fundamentals instead of replacing them. ㅤ What We CoverHow AI changed buyer behavior: Buyers now research solutions through ChatGPT and Claude, not just Google Search. Megan explains why the companies getting cited inside those tools are the ones with clear positioning and real brand investment, not the ones with the slickest automation.Losing the plot with AI agents: Most teams are obsessed with building AI workflows to look efficient. Megan argues this misses what actually matters: positioning, differentiated messaging, and brand strategy. The irony is that fundamentals matter more now, not less.Why positioning may already be stale: The messaging you wrote two years ago might not be relevant to your buyer's new reality. Megan recommends re-running your strategic narrative and going back to talk to buyers about pain and benefit.Content is about what you say: With AI commoditizing production, the differentiator is your point of view, not how fast you make content. Megan frames content strategy around context and a unique perspective.The marketing maturity model: Megan describes the assessment Refine Labs uses to look across sales, marketing, and customer success and find the gaps in a go-to-market engine. It's how a marketing leader earns a real seat at the table.Measuring brand for skeptical CFOs: Self-reported attribution and last-touch both have a place, but neither measures brand on its own. Megan shares how share of search trends give executives a concrete signal over time.The LinkedIn incrementality test: One client split its target account list in half, ran brand ads to one group for 120 days, and held the other back. The treated group produced 60% more qualified pipeline, which won over a skeptical C-suite.The CFO lens on spend: Megan recommends looking at total all-in marketing investment against new business acquisition to find real ROI and contribution margin, rather than fighting over individual channels.The biggest wasted budget: Google Ads spend with no proper conversion tracking, and LinkedIn campaigns running direct response when they should be brand awareness. Megan notes she often drives the same or better results for 30% less by auditing Google.Founder brand is an asset, not a strategy: Refine Labs was built on its founder's voice and never diversified. When he left, the company felt it. Megan explains why one voice should be part of a broader strategy, never the whole thing.What to outsource versus hire: Keep narrative, pricing, product marketing, and content strategy in-house. Outsource specific channel expertise like paid media, where outside partners bring data and benchmarks you can't see alone.Where video is winning: Megan is seeing more clients invest in YouTube, CTV, and out-of-home so buyers can watch something, feel it, and see people like them using the product. ㅤ Resources MentionedRefine Labs: Megan's agency and the source of the frameworks and experiments discussed throughout the episode.ChatGPT: Named as one of the LLMs buyers now use to research and discover solutions.Claude: Cited alongside ChatGPT as part of how B2B buyers now look for products.Google Ads: Discussed as the most directly attributable paid channel and a common source of wasted spend without proper conversion tracking.LinkedIn: The channel behind the incrementality experiment and Megan's argument for brand awareness over direct response.HubSpot: Referenced as an internal system that needs proper conversion tracking tied to Google Ads.Salesforce: Named alongside HubSpot as a system for connecting conversions to pipeline.Meta: Suggested as a paid social alternative to LinkedIn depending on the audience.Reddit: Mentioned as another paid social channel companies overlook in favor of LinkedIn.YouTube: Cited as a video channel where Megan is seeing strong results for clients. ㅤ Safe Doesn't Scale is hosted by David Walsh, founder of Limelight. New episodes drop weekly.

    Everyone's Obsessed With the Wrong Thing in AI Marketing (with Megan Bowen from Refine Labs) | Ep. 17
  7. 18 June

    The Biggest Risk on LinkedIn Isn't What You Post (with Dani Markovits from Shake Content) | Ep. 16

    Everyone can make content now, so everyone does. The result is a feed full of posts that read like they came out of the same prompt: same hooks, same tidy takeaways, same manufactured vulnerability. The part nobody mentions is that LinkedIn has quietly started grading the person behind the post, not just the words in it. ㅤ David Walsh, founder of Limelight, sat down with Dani Markovits, who spent four years building LinkedIn's creator program before joining Shake Content as chief commercial officer. They get into what actually works on LinkedIn now, why the same post lands differently depending on who hits publish, and how to keep a posting habit from eating your week. You'll leave with a system that fits in an hour. ㅤ Dani brings the view from inside the platform: how the creator team thought about the feed, what the new algorithm changes reward, and why he almost turned down the agency job. He also talks through the moment early on when he sat down to post and could not think of a single thing to say. ㅤ Guest BioDani Markovits is chief commercial officer at Shake Content, a LinkedIn-first B2B marketing agency based in London. Before joining Shake, he spent four years at LinkedIn as one of the first members of its creator team, working across Europe with hundreds of executives, founders, and athletes as the platform reshaped itself from a job board into a place people return to daily. He's now building content programs for B2B tech companies, professional services firms, and a growing roster of athletes. Somewhere along the way he picked up the nickname "the LinkedIn whisperer," which he says he wasn't a fan of at first and has since decided to own. He'll also admit, freely, that he still struggles with video. ㅤ What We CoverWhy he left a big-name role: Dani explains why he turned down Shake at first, thinking it was too small a move from LinkedIn, and what changed his mind. He frames it as a mutual risk: he bet on them, they bet on him.What a "creator" actually is: He never felt like a creator, because he pictured polished YouTube video. His reframe: if you share your thoughts and expertise, you're already one, and it doesn't have to be your whole career.The authenticity premium: With AI slop flooding the feed, the content that stands out is the stuff only you can write. David adds the contextual-storytelling angle: "I was actually there, I actually did this."Why slop is an opportunity: Dani argues the flood of generic content is good news for anyone willing to play the long game without cutting corners. More noise raises the payoff for real insight.Beating the blank page: His fix for "what do I post about" is to stop hunting for original ideas. Look at your calendar, who you spoke to last week, what you're reading, then bring it back to your own lens.The one-hour-a-week system: Block an hour. Get one post out without overthinking it. Spend the rest leaving real comments and sending ten connection requests to ICPs, prospects, and people you respect.Sweated posts underperform: The posts you edit five times and sleep on are often not the ones that work. A quick thought or something funny frequently does better.Who's posting now matters: Under LinkedIn's new changes, the platform weighs the author as much as the content. The same post about running a business does better from someone who has actually run one.Comments as a strategy: A good comment is content in itself, and LinkedIn is prioritizing it. Many people now get more growth from commenting than from posting.The anti-pitch-slap: Cold DMs work, but engaging with someone's content for two weeks before reaching out makes a positive reply roughly ten times more likely.Boring industries win: The duller the field, the more room to stand out. A tax accountant who posts has far less competition than yet another marketing voice.The real risk: Dani's closing line is that the biggest risk on LinkedIn isn't picking the wrong thing to post. It's not posting at all. ㅤ Resources MentionedShake Content: Dani's LinkedIn-only agency, discussed throughout as the home for the founder and executive content programs he now runs.LinkedIn: the platform at the center of the conversation, where Dani spent four years on the creator team before advising clients on it.SaaStock: the SaaS conference where David first met Shake's CEO, James, who later pitched him on the business. ㅤ Safe Doesn't Scale is hosted by David Walsh, founder of Limelight. New episodes drop weekly.

    The Biggest Risk on LinkedIn Isn't What You Post (with Dani Markovits from Shake Content) | Ep. 16
  8. 11 June

    Why Most B2B Influencer Reports Don't Survive a CFO Meeting (with Will Beech and Chris Peters from Moon at Dawn) | Ep. 15

    Most B2B influencer reports celebrate impressions and call it a campaign. The CFO signs off because the spend sits inside marketing. Nobody walks into the renewal confident about what the money actually bought. ㅤ On this episode, David Walsh, founder of Limelight, talks with Will Beech and Chris Peters, co-founders of Moon at Dawn, a tech-first B2B influencer agency built around outcome-based pricing. They get into ICP resonance, attribution, and what it takes to put guarantees on an influencer campaign. ㅤ Guest BioWill Beech is co-founder of Moon at Dawn, leading client strategy and product. He spent eight years in B2B influence at Onalytica, working on programs for Siemens, IBM and AWS. ㅤ Chris Peters is co-founder of Moon at Dawn, with 15 years in agency land, most recently as B2B Global Client Lead at Wavemaker (WPP). He writes the B2B Excellence newsletter. ㅤ What We CoverWho Moon at Dawn sells to: The three personas they target: enterprise brands, mature brands already running programs, and mid-market companies. Each measures success differently.Influencer, not creator: Why they treat "influencer" as an umbrella covering subject matter experts and executives, which lands better with enterprise buyers.What outcome-based pricing means: Guarantees on impressions against ICP accounts, frequency, and lead volume at a cost competitive with paid media.The $20K floor: Around $20K for an outcome-priced engagement, $10K to $15K for an organic-only pilot. Below that, the data won't support a guarantee.Resonance over reach: Why a single CEO like is happenstance, and how depth across the full buying committee is a better signal of commercial impact.The Cyber Things campaign: A cybersecurity pilot tied to the Stranger Things finale that doubled the previous agency's metrics through ICP-based selection.Aurora Scouts: The platform that surfaces collaboration opportunities, flags risk signals, and alerts the account lead when a client gets acquired.The agency of the future: Chris on why 30 people and £3M turnover is the dead zone, and why the next acquisition cycle rewards service businesses with real tech inside. ㅤResources MentionedOnalytica: where Will built his B2B influencer career.B2B Excellence newsletter: Chris's Substack, recommended by David on air.Riverside and Gamma: cited as examples of mature brands running sophisticated programs. ㅤ Safe Doesn't Scale is hosted by David Walsh, founder of Limelight. New episodes drop weekly.

    Why Most B2B Influencer Reports Don't Survive a CFO Meeting (with Will Beech and Chris Peters from Moon at Dawn) | Ep. 15

Trailer

About

"What's the ROI?" Those three words kill more creative marketing ideas than bad execution ever will. Not here. Safe Doesn't Scale is a weekly podcast for marketing and growth leaders. We’ll be interviewing Heads of Marketing, Unicorn Founders, and Revenue Leaders at B2B companies to prove that the riskiest marketing campaigns drive the biggest returns. While brands are burning $500K on LinkedIn ads that are generating zero demos, there’s someone out there who closed a $2M deal they sourced from a meme. Host David Walsh, Founder of Limelight, breaks down real examples from brands spending less and converting more by leaning into creator-led growth, unconventional distribution, and campaigns that make traditional marketers panic. You’ll learn: How growth leaders sell “unsafe” ideas to the C-suite How to attribute sales pipeline to content, creators, and social signals Why the campaigns that feel uncomfortable often drive the most revenue No e-book downloads. No buzzwords. This show is for marketers with a chip on their shoulder who are tired of playing it safe. We celebrate the campaigns that make legal sweat and sales teams crush quota. Because marketers who don't take risks won't exist in 2027.