Clay Ads is live! We've been using it for 7 months and dropped our LinkedIn CPL from $250 to $25. Meta went from being unusable to a $10 CPL. I'm so pumped about this product 🔥 Here’s why it matters right now. CMOs are moving more marketing budget into paid ads as other channels get noisier, but are struggling to keep spend efficient. Pouring more budget into existing ad campaigns hits diminishing returns, because you're paying to reach the same fixed pool of people over and over. Spending efficiently means growing the pool of fresh, qualified contacts faster than you grow the budget. Meaning, increasing ROAS comes down to two things: 1. Who you're targeting. Clay Ads builds audiences from your CRM, data warehouse, and 200+ data providers, filtered down to exactly who should see the ad. They update automatically as your pipeline moves, so customers and open opportunities drop out on their own. 2. Whether they actually see it. Most B2B companies only have work emails, and platforms like Meta and Google can barely match on those. Clay uses its data marketplace to find up to 3 hashed personal emails per contact, so your ads reach buyers where they actually spend time: Instagram, Google, LinkedIn, and beyond. Get match rates as high as 95% on LinkedIn, 85% on Google, and 70% on Meta. Your ad campaigns are only as effective as the people who see them. Now you can target the right accounts (the ones you actually care about!) and the right titles across every channel, without wasting money on the rest. Clay Ads is available today on Growth and Enterprise plans with unlimited ad audiences. On August 13th, we're doing a livestream to walk through how we run our own paid media on Clay Ads. Join us here: clay.link/WpaHxGO | 35 comments on LinkedIn

Here's how we fixed inboxes running a 40% sender bounce rate. Not a bounce because the address was bad. The addresses were validated. Google was refusing to let us send the email at all. The first time I saw this, I couldn't wrap my mind around what was happening because the lead list was validated. We revalidated it again and it was still happening. After a bunch of trial and error, we found the Google had fingerprinted the text we were sending and wasn't letting the message even get sent. We spintaxed the subject line and the unsubscribe language with over 20 variations. We already had a little spintax in the body. The sender bounces went away immediately. I want to be honest that a year ago I was the guy telling people spintax didn't matter. We were using so much AI in the message that I figured the variation was already there. I was wrong about that. What changed my mind is that the fix worked in isolation. Same domains. Same inboxes. Same offer. Same list. Only the subject line and unsubscribe language changed, and a 40% sender bounce went to zero. What we do now internally, and what I would tell anyone running real volume: 20 variations in the subject line. 20 in the unsubscribe language. Variations in the signature block. Three or four versions of the overall body, and three or four options every five words or so. We have Claude skills that write all of it, because writing 20 subject line variants by hand is miserable. If you are reading this thinking "his big tip was spintax," I know. But the amount of spintax you think you need should 5-10x for the health of your domains. | 18 comments on LinkedIn

We've worked with companies ranging from 1-person startups to the Fortune 500. Here are some of the most valuable GTM engineering use cases for each: Early-stage startups ($0-$10M): Most of the businesses we've worked with at this size care almost exclusively about generating top-of-funnel. Real "move fast and break things" vibes, with less of an immediate concern over data hygiene. Because of this, the two projects we historically have built most for them are outbound projects (automated or semi-automated outbound) and list building. Late-stage startups ($10M-$50M): At this stage, companies begin to care about building real systems instead of the "run and gun" approach that got them out of the trenches. Many of them ask us for help with ICP definition based on past closed-won, more refined approaches to outbound (not just running one campaign to everyone), and account scoring so they know what prospects their sales team should prioritize. We also start to see inbound lead enrichment here for companies that get a lot of leads through marketing. Growth stage companies ($50M-$200M): This is when more advanced use cases start popping up. Systematized signals (as opposed to ad hoc) are a BIG one now that sales teams have quarterly goals that may not be fully attained from prior initiatives. They also tend to start with some basic CRM hygiene at this stage, mostly just to help with active deals so they don't slip from data gaps. We also start to see territory carving and automatic lead assignment to reps. Midmarket ($200M-$1B): This is when CRM hygiene and enrichment becomes a big use case. We also start to see big changes with outbound workflows -- companies at this size sometimes have hundreds of reps, which changes the amount of human effort you're able to put into outreach. GTM Copilot workflows often replace the automated outbound campaigns we run for clients when they get to this stage. Renewal intelligence (cross-sells, whitespace, etc) is also big here as companies look to retain and expand existing logos. Enterprise companies ($1B+): Frankly, most of the enterprise companies that we've worked with need ALL of the aforementioned use cases, and in a compliant manner (shoutout Clay for SOC-2). We see CRM hygiene, the marketing -> sales handshake, and broad outbound systems being the three biggest headaches for companies at this size. Given the scale already achieved in their CRMs, outbound tools, and data warehouses, this requires a real surgeon of a GTME team to unpack and rebuild (which we pride ourselves in having), as well as a meticulously scoped plan. On the positive side, the value return for GTME work in these orgs is the largest by FAR of all of these groups. Even small tweaks in outbound or data processes can yield millions in savings and pipeline. Curious to hear people's experiences here! Of course, we've seen exceptions to the above, but wanted to give those new to the space a general idea. | 19 comments on LinkedIn

Bootprint 🥾 (prev Clay Bootcamp) is launching 6 life-changing scholarships... … to help the winners get jobs and promotions as GTM Engineers … and kick open doors they couldn’t open alone. – With our renaming we’ve doubled-down on helping people get jobs as GTM Engineers, and future-proof their careers. And we’re so grateful to our sponsors & our partners: Clay, Deepline, Attio, OutboundSync, EmailBison, Understory, FastForward, RevenueHoop, and Girls Who Clay. … who not only support these scholarships… … but many of whom are covering tool costs for ALL future Bootprint students. 🏆 For scholarships we will be selecting: • 2 Women scholars (follow Girls Who Clay, RevenueHoop 🙋‍♀️) • 2 Americas scholars 🇺🇸🇨🇦🌎 (North/South/Central!) • 2 Wildcard scholars from ANYWHERE! (e.g. APAC, EMEA, etc.) 🏆 WHAT'S INCLUDED in the program: 1️⃣ Lifetime access to the “Harvard” of GTME networks People often say this network, by itself, is worth the price of admission. 2️⃣ [REVAMPED] 4 impactful edu programs: 🎓 C4C - Clay for Careers 🎓 C3 - Claude Code Cohorts 🎓 CRM - CRM+RevOps Masterclass 🎓 CEM - Cold Email Machine ft. Claude Code 3️⃣ 12 weeks of 1-1 coaching on real projects. Can be used before or during a new or current job. Dozens of incredible coaches and mentors, matched at just the right time, support our students here. 4️⃣ [NEW] Job finding (optional) + career guidance Many joiners want to keep doubling down on creating value for their current job, but for those seeking… Many grads have earned jobs at places like Understory, The Kiln, Octave, Clay, ElevenLabs, and many more. Bharat Devana helps you skip the line with job matching. And our LinkedIn King/Queen-maker Lab helps you attract lots of opportunities. 5️⃣ [NEW] $3,500+ in core GTM Engineering tools So you don’t worry about extra expenses as you learn. — 💪 Battle hardened with 350+ students over 3 years: • Grads hired by Clay, ElevenLabs, Understory, The Kiln, Octave, and many more • Clay’s trained 40+ employees with us • Vibrant community full givers with success stories ❓Apply if: • You’re ready to “lace up” your boots and put in the work (10+ hours/week) • You’re a giver and want pay it forward as you win • You want to become a GTME & change your life ⏰ Deadline to apply: Tuesday, August 18. ⏰ Program start date: Tuesday, September 8. — ⏩🚨 How to apply: • Like this post • Comment “Boots” below (must be connected; add me!) • Repost this post • Fill out the scholarship form we send! (must be connected) Best of luck! 🥾 | 153 comments on LinkedIn

How to double your outbound output without doubling your headcount. Most sales teams run 2 SDRs across email, LinkedIn, and cold calls. Their output: → 2,200 emails/month (100/day) → 200 LinkedIn connects (20/day) → 1,100 dials (50/day) Monthly totals: → 4,400 emails → 400 LinkedIn connects → 2,200 cold calls Now replace one SDR with a GTM Engineer and delegate by skillset. → GTME owns email and LinkedIn. → SDR goes all in on cold calls. GTME sends 20k+ emails and manages LinkedIn automation through both profiles. SDR gets all that time back to dial through a parallel dialer (200 dials/day). Monthly output for GTME + SDR: → 20,000 emails → 400 LinkedIn connects → 4,400 cold calls Here's when you know it's time: → Your SDRs are building their own lists between dials → Your deliverability is inconsistent and nobody's monitoring it → You've got no warmup process, no inbox rotation, no domain strategy If two of those are true, your next hire should be a person who can build systems for you sales reps. What's your team's current split across channels?

PSA- Clay's “Find Companies” search has gotten WAY better. For those who have been using the tool for a little while, you know that Clay’s previous filters were essentially just LinkedIn’s filters, so finding the exact list of companies/people you wanted was pretty tough. Now, they’ve unified all of the providers into a singular list of filters, which makes this list-building function significantly easier. They also have natural language search, ai-code-pasteable formats for any filters you create, and you can even run boolean/OR statements in your filters to run more robust searches. You can even create custom industries if you don’t like what you see in the default options (for example, “Crypto”) and Clay will create the list for you This may seem small, but for people in the product, it’s a massive improvement. Would love to hear if people have been using this, and if so, what your thoughts are! | 17 comments on LinkedIn

I believe Buzz is a bigger moment in AI than when OpenClaw first came out. Every frontier lab is going to copy this. But I don’t think they win unless they embrace open models. What is Buzz? In simple terms, it’s an open-source version of Slack, OpenClaw and Herd. It's built by Jack Dorsey who also created Twitter. Buzz lets you interact with agents on top of any harness: Claude Code, Codex, Hermes, whatever you use, while collaborating with your team in a familiar Slack-style workspace. Here’s what I really like: → Shared compute: owners can share local compute with members, so a community can essentially pool its hardware and run larger local models together → You can use Hermes, Codex and Claude Code in one Slack-style workspace → Delegating work to agents feels great. Tag an agent or two, and it replies in a thread with status updates as it works through the task It’s the first multiplayer agent harness. | 17 comments on LinkedIn

The anatomy of a 12% reply rate email: From trigger → problem → solution 1. Trigger Your trigger is everything. Make it relevant and tied to an issue you solve: - Job changes - Company hiring - New product launches The key is to find the trigger that leads to the issue you solve. 2. Implication Then tie that trigger event to an issue that you solve for. Make sure to use hedge words Thought that… Imagine that… Curious if… Plant the seed of the issue before calling it out. You want it top of mind for your prospects. 3. Pain Next call out the surface-level pain. Not every pain they could have, just the biggest. Literally whatever issue YOUR customers face. Don't use generic pains like time & money. It has to be relevant to your ICP & product. Be specififc. 4. Second-order consequence A surface-level pain is not enough. You have to go deeper and show what happens if nothing changes. How does that issue affect a KPI your buyer cares about. That’s what will get you responses. 5. Social proof Then show how you helped someone like them. Not all social proof is equal here’s how to rank it: Big name company + same industry + big outcome Same industry + Big outcome Big outcome If you nail all 3, it builds more trust than 1000 words. 6. Solution Now we can move into our actual solution. We want to make it simple and build curiosity. The solution is just the vehicle to get them to the outcome from before. So a great solution has to have good social proof before it. It should be 1 sentence max. 7. Soft CTA The last part is your CTA. Our 1 goal is to get a response and start a conversation. Make it as easy as possible to respond to. When you get the first response, you’re in. We’ve taught the TIPSSSS framework to over 400 reps 1-on-1. The best have seen 12% reply rates on automated sequences. What would you change in this email? 👇 P.S Save this post and send it to an AE who needs to see it! | 29 comments on LinkedIn

Smartlead analyzed 850M+ cold emails sent in 2026. Here's how the top 1% actually run cold email: 1. Audience > list ↳ The best senders get a reply every 38 contacts. The average needs 135. → Build audiences you can describe in one sentence and why you're emailing them in the first place. If you do this, you already have the copy. 2. Short emails > long ones ↳ Under 50 words gets the most replies. Most people write around 98. → Cut the first email in half. Say one clear thing. Have a documented, clear, value prop in plain, simple English, optimized for 5th-grade language. 3. First hour > day seven ↳ If you get responses, you'll get them within the first 3 hours. → Follow up in the first day or two (not a week later) to maximize conversion. 4. Morning > afternoon ↳ Sending during the morning hours gives you better results than sending in the afternoon. ↳ There is no magic day during the week, but weekdays perform better than weekends. → Send early. Same effort, more replies. 5. A sequence > one email Most replies come from the first email, but sending only one is not enough: ↳ Email 1: a personal opener and something useful ↳ Email 2: a different angle, a couple of days later ↳ Email 3: a short last try → Keep them short and refer back to the first email + your unique insights. 6. Clean data > volume ↳ The average bounce rate is 1.54%, and many people's bounce rates are 3%+. → Set up SPF, DKIM, and DMARC → Keep bounce under 2% → Diversify instead of relying on one single provider (Scaledmail, Inboxkit, etc) → 2 mailbox/domain → minimum 3 weeks of warmup → no links, images, GIFs in the emails, plain, simple text → Add mailboxes instead of pushing one harder, and invest in GOOD data. 7. Offer first > ask first ❌ "Can we hop on a call tomorrow at 2pm?" ✅ "Can I send over to [specific problem]?" → Let them say yes first. One thing the report doesn't articulate clearly, but I'd like to:: #1 Data > copy. It's easy to obsess over the copy: ↳ Can AI write a clever opener? ↳ Can AI mention their funding round and last LinkedIn post? ↳ Can AI stuff in as many personalized tokens as possible? And lose sight of the only question that matters: WHY ARE WE EMAILING THIS PERSON IN THE FIRST PLACE? When you actually know who you're writing to, that answer is obvious, and the copy takes care of itself. You just can't get there without the data. Think of them as an audience that can be described well, not as a list. #2 Own data > rented data. Buying funding or hiring lists works for a while, but everyone buys the same ones, so there is literally zero impact or unique approach in this. The data you collect yourself will be your durable strategy, because no one else has it. #3 Data 1st > Activation 2nd Get the data right, then automate it, then activate through GTM. Most teams do it backwards, and that’s why your funnel isn’t working. Comment "REPORT" and I'll send it over (we need to be connected so I can DM you.) | 121 comments on LinkedIn

I don’t know why Founders are sleeping on this crazy growth strategy. Word of mouth and referrals are good growth strategies. But you can't forecast them. Right! And I bet you don’t want to build a pipeline around someone else's dinner conversation. To grow rapidly and consistently, you need a FUNNEL. Not some Russell Brunson, 47-step webinar monstrosity. Just a simple one would do a great job. Here's what a basic inbound funnel looks like: Step 1: Your content (LinkedIn, Meta, whatever) gets someone to click You have to be embarrassingly specific with your CTA. Not "check out the video." You say: "Watch this video. If you like what you see, book a call with us." Because if you just tell someone to watch, broski, they'll watch and bounce. Step 2: They land on your VSL page This is where 80% selling happens. Your VSL needs social proof, problem agitation, and benefits broken into stages. Don't promise "you'll lose 90 lbs." Break it down. "At 14 days, this happens. At 30, this. At 90, this." The staged breakdown builds belief way faster than one big claim. Step 3: Application form You qualify them before they ever touch your calendar. Three questions worth asking: where are you right now, what have you already tried, and what does winning actually look like. Step 4: They book the call Booking doesn't mean showing up. Show rate is where most funnels quietly bleed out. Three things you can do. One, send an immediate confirmation. Two, on the thank you page, literally spell it out: "Go to your inbox. Click confirm. Not in your primary? Check spam." Three, 24 hours before the call, send a personal message. Pull what they wrote in the form. "Hey, we'll talk about how you can hit [their goal] in 13 minutes. See you at [time]." That's it. That's one funnel. You can build different versions for outbound, referrals, partnerships, whatever. The point is that every single stage has a job, and if any stage is just sitting there looking pretty, the whole thing falls apart. So genuine question: how many meetings did you book last month? | 14 comments on LinkedIn

Harvard Business Review just leaked one of the hottest AI jobs that didn't even exist a year ago. AI managers are going to be the standard job title within 18 months. Now with these 100+ resources, you can learn it for free: Multiple incubators, notable consulting experts, and Harvard Business Review named the Agent Manager the job of the year. The AI agent job market is expected to reach billions. This job title didn't even exist one year ago. And the AI agent manager is going to be the standard job title within 18 months. AI usage is shifting from chat-based to action and outcome-based. With this, it's increasingly important that these AI agents drive the right business outcomes by: - monitoring their performance, - optimizing workflows, and how agents interact with each other and team members, - diagnosing when agents are messing up, - and measuring the AI agent ROI. Your job is not going to disappear if you: - have domain expertise, - can apply your creative skills for problem-solving, - and make a calibrated judgement (what AI cannot do!). So to help you, I've mapped 100+ resources that take you from zero to hired, across 4 tracks. Most are TOTALLY free. Here are 9, grouped by where you are today: If you're starting from zero: ↳ 1. Anthropic, Building Effective Agents. The one guide everyone should read first. ↳ 2. Microsoft, AI Agents for Beginners. 18 free lessons on GitHub. ↳ 3. Hugging Face, AI Agents Course. Hands-on, free certificate. If you want to manage agents, not build them: ↳ 4. HBR, The Agent Manager. What the job actually is. ↳ 5. Agentic AI: Reinvent Business, Work, and Life, by Pascal BORNET. A plain-English view for leaders. ↳ 6. McKinsey & Company, Rethink talent for agentic AI. How teams are restructuring. If you want to build your first agent: ↳ 7. AI Agents in LangGraph, DeepLearning.AI with Harrison Chase. Build one this weekend. ↳ 8. Multi AI Agent Systems with crewAI, DeepLearning with João (Joe) Moura. Make agents work together. ↳ 9. Mastering Agentic AI Certification, by Aishwarya Srinivasan. The certificate that ends up on your profile. The full list has 100+ resources, sorted by level and format, plus the exact order to take them in. Comment "LEARN" and I'll send it your way. (Make sure we're connected so I can DM you) Repost this if someone in your network is trying to figure out where to start. ♻️ P.S. We're improving our agents and agentic workflows. If you have a CRM with a team of 5 sales reps (mix of AE, BDRs), I'd love to chat! | 59 comments on LinkedIn

I've spent two years posting about MCP servers and API workflows. I've probably made 6 posts about the new Clay CLI in the last week. Then Clay Cup put me in a room with 20 minutes, four connected puzzles, and no agents. Competing against the best builders in the world. Venues, local data, four puzzles that fed into each other. Eliminations at the 10 and 20 minute marks. No Claude Code, no MCP, no "let me just script this real quick." Just Clay and whatever muscle memory I still had. The muscle memory held. Barely... On to the Round of 8. Congrats to everyone who both competed and advanced. Thanks to Clay for building something that makes 300+ people want to prove they can do this under a clock. Owen Ciha, Munnawar Hashim, appreciate the long nights to make this possible. Andrew Quach and Siya Verma🦄✨thanks for looking after us! | 20 comments on LinkedIn

The best part of cold email is everything before you press send. 😂 Buying domains feels like progress. Loading up 5,000 prospects feels like progress. Watching the campaign go live feels like you're about to print meetings. Then... nothing. So you change the subject line. Then the CTA. Then the tool. Then you convince yourself deliverability is the problem. A few weeks later you finally ask the question you should've asked on day one… Would these people even care about what I'm offering? That's usually where the real work starts. Cold email isn't hard because of the emails. It's hard because good targeting and a great offer can't be automated. What's the lesson that cost you the most in outbound? | 152 comments on LinkedIn

If you still need five tools to prospect in 2026, you're working way too hard. 19 B2B plays you can steal today, all from one search 👇 Finding companies was never the hard part. Knowing which ones, why now, and what to say is. I've been running Prospeo inside Claude. These are the most relevant 19 filters for any B2B angle - you can get all these in one search: TARGET BY HOW THEY SELL 1/ Product-led: Business Model = self-serve + free trial → Claude: outreach that leads with time-to-value. 2/ Sales-led: Revenue Model = sales-led + demo → Claude : write for a demo buyer, not a trial. 3/ Sells to your buyer: sells to [buyer title] → Claude: partnership angle, same audience. 4/ Lookalikes: Company Lookalike = [best account's domain] → Claude: open with why you fit teams like this. TIME IT WITH A SIGNAL 5/ New leader: Exec Events = new CEO/VP, last 60 days → Claude: short note tied to their first 90 days. 6/ Just raised: Funding = last 90 days → Claude: tie their raise goal to your value prop. 7/ News-based: Company News = launch/expansion/M&A → Claude: reference it, don't pitch the product. 8/ Traffic up: Website Traffic = trending up → Claude: scale-up angle. 9/ Traffic down: Website Traffic = trending down → Claude: growth-audit offer. QUALIFY BY WHAT THEY ARE 10/ Compliance-driven: Certifications = SOC2 / ISO / GDPR / HIPAA. 11/ VC-backed growth: venture-backed + revenue + headcount growth. 12/ Has API / uses AI: AI Attributes = has API → Claude: integration pitch. 13/ Team in one country: Headcount by Location = [range] in [country] → Claude: write for that market. 14/ Competitve takedown: Competitive Intel = mentions [competitor] → Clude: switch pitch, no trash-talk. 15/ Lists a brand you know: Key Customers = has [brand] → Claude: reference the shared connection. GO DEEP ON THE ACCOUNT 16/ Ranks for a keyword: Google Discovery = [keyword] → Claude: niche-aware opener. 17/ Says an exact phrase: Company Keywords = [phrase] → Claude: their own words back at them. 18/ Hiring now: careers page hiring [role] → Claude: fill-that-gap angle. 19/ Whole buying committee: Smart Job Title = expand [title] → Claude: one message per role (buyer, user, blocker). How to run it with Claude: 1. Talk to it: Prospeo MCP inside Claude, "build a list of [your ICP]" — no exports 2. Check before you spend: Search Analytics, export only the segment worth the credits. 3. Hand off clean: push a tiered list to your CRM, best-fit accounts get the call first. Don't pitch on the first touch. Open with something useful, then iterate. Do you find these useful? P.S. If I earned your trust, I'll set you up on Prospeo and automate all your TAM sourcing at a better price than Apollo. Book a time with me. | 94 comments on LinkedIn

Today Clay is launching Account Research Agents! This is the first of a series of always-on GTM agents we'll be releasing this summer. Here's how it works: these agents monitor all your accounts, reading new call recordings, emails, signals, and warehouse data. They can synthesize and write key updates to your CRM or data warehouse automatically, such as who the champion is, what stage the deal's in, what the risks are, why you lost. No more harassing your sales team to update Salesforce. In closed beta, teams used the agent to flag high-usage customers for expansion, revive closed-lost deals with the real loss reasons, and quickly ramp new reps with full account history. Account Research Agents are now available on Enterprise, Growth, and Launch plans. Check it out here: clay.link/t3GoXtv | 20 comments on LinkedIn

We trained an AI on 3,019,536 emails and 2,355,916 linkedin positive replies generated by our customers at lemlist (real numbers). Unlike ChatGPT or Claude, it's specialized in outbound. And it doesn't need a separate subscription. We call it lemAgent (for now). Here's the idea 👇 Today, launching a campaign means jumping between 3 tabs: ChatGPT to write, a lead tool to find people, lemlist to build the sequence. An hour later you've got a half-built campaign and a lead list that's almost right. lemAgent kills that: one conversation. Say you're an SDR at Pennylane (lemlist's cutomer) and selling to CPA firms. You tell lemAgent: "Find accounting firms with 10-50 employees and build me a campaign to reach their partners." From that one sentence, it: → Finds the accounts: searches our 650M database and pulls accounting firms in that size range and your region, with the partners' names and contact info → Ranks them by buying intent: like firms with multiple open position for junior accountants (drown in manual work), or firms that visited your website, or even firms with recent m&a activities. → Builds the full sequence: email + LinkedIn, right timing, right steps → Personalizes every message: references the firm's recent hiring or growth, not a generic "Hi {{firstName}}, I hope this email finds you well" One important thing about this agent: it's not autopilot. Because we don't think autopilot works for outbound today. So we built a copilot, eg "human in the loop" It asks you questions about your business, your USPs, your ICP to actually understand before building a campaign that can generate meetings and revenue. lemAgent builds the campaign, but you still run the show. Nothing goes out without your sign-off, so it doesn't harm your brand. A few weeks into the beta, 10% of our daily active users are already on it. Now the fun part: we need help to find a better name than lemAgent. We were thinking something spatial as "lem" stands for Lunar Excursion Module, but ideally human, like "Ariane." Drop your idea in the comments. If we pick yours, I'll send a nice gift 🎁 | 29 comments on LinkedIn

I think the best cold email campaigns have 3 things in common. A campaign that drives ROI doesn't need all 3, but each one of these points drives significantly more replies. Ever since we started doing free tests for companies to work with us, we've launched quite possibly 3-5x more campaigns with different offers than we previously would. Before, I thought I could tell from one look if a campaign was going to work, now, I'm basically never surprised. I've tried to put my finger on what makes a great campaign versus others and this is what I've been thinking. 1. Great campaigns communicate the promise to either help companies make more money, save 90% of time, save 90% of costs in a way that the recipient didn't know about with a next step that is easy to say yes to. The best type of offer to convert cold traffic is an offer that helps people make more money for sure. If you can, always lead with that. If you can't, the time savings / cost savings you can provide has to be orders of magnitude higher. When Instantly.ai first released a sequencer with unlimited inboxes where everyone else was charging per mailbox, the community went crazy. And that was maybe a 99% cost reduction depending on your scale. Save Time/Save Money offers have to be orders of magnitude better than what they are doing today. 2. Great campaigns communicate a mechanism that the recipient didn't know existed. I think the increase in amount of outbound messages people are receiving is maybe responsible for 20% in the decrease in outbound effectiveness. (I'll make another post demonstrating this). The bigger driver of the decrease in outbound effectiveness is that prospects have access to more information than ever before. They know that performance outbound lead gen, bookkeeping, HR management, payroll automation, cyber security, recruiting, are all things they can get access to. And when they're ready they will find you. They don't know that - Trigify.io can find customers talking about their exact problem - That negative google reviews can be removed with a proper appeal process - That Forbes is more pay to play than you might think - That the Donut can target newsletter subscribers by income/interest - That Starbridge can get them a list of governments that have open RFPs. So the second thing great outbound campaigns do is present a mechanism they didn't know was possible. 3. The best outbound campaigns explain the value in taking the next step AND the next step is easy. It costs Young Dental $200-$300 to get a dentist to request a sample through Facebook ads. With us, it costs ~$40. Their campaigns if anything are a time saving / reducing risk of a bad patient experience type offer. Not make more money. And these dentists don't know that these products exist. But the call to action being to get a free sample set sent to their office is such a no brainer that they get tons of responses. Last part in the comments, this was too long | 22 comments on LinkedIn

Last week Clay released an API after all of the smartest Clay pundits in the world told me they’d absolutely never ever do it. What does THAT tell you about the state of the world? My Take: You’re an idiot if you’re building UI for people. The argument, made to me by INCREDIBLY smart people, was that a public API would hurt Clay's business by making it less sticky. Clay was sticky BECAUSE workflows lived in the UI. Let people build on an API, and those workflows go elsewhere. It also devalues Clay's vendor relationships, because you no longer have to log into Clay to access that data. Which was their other "moat." Now, it's possible that 100% of the revenue Clay actually cares about is Enterprise, and that revenue will be VERY slow to move onto an API/CLI. If that's the case, Clay may have done this to get UGC going again. UGC comes from the SMB segment, and you can't kill stickiness that's already dead there. If that's the play, tip of the hat to Kareem and the Clay team. Wise move. But even so: Clay just did something that, if you asked Claude whether they should do it, Claude would say "no, that kills your moats and your stickiness." That's a HUGE statement about the state of the world today. #1: I love that at least ONE other CEO isn't blindly taking direction from Claude. #2: This way of doing things is in such high demand that a $5B darling unicorn is willing to satisfy it, even though it's very not good for the stickiness of their business. What am I getting at? If I were starting a new business today (and I am)... I would service this use case and do nothing else (which I am). UI for humans is dead. Long live APIs. 🏴‍☠️ | 71 comments on LinkedIn

1 in 4 of your posts should convert. The other 3 just warm the room. Here is the 4-tier content score we run at Growth Today: In the same way we tier an outbound list, we also consider the same logic in the content strategy. Each content piece has a job, and the tier a reader lands in decides the message they get later. Here is the 4-tier content score: TIER 3, BoF, convert ↳ a real asset: a lead magnet, a case study result, a teardown with proof ↳ With the highest intent, the people who engage here raised their hands ↳ The DM it earns: reach out within a day, name the thing they grabbed, or if it’s not relevant, just keep your outreach short + sweet TIER 2, MoF, prove ↳ How you think, how you solve a problem, a contrarian take with a method ↳ This is the workhorse; it makes the reader think you know your stuff ↳ The DM it earns: a personal note that references their comment, no pitch yet TIER 1, ToF, reach ↳ a story or a stat that attracts and pulls a broader audience ↳ grows the top, new followers, future pipeline ↳ The DM it earns: none yet, a like is not a buying signal TIER 0, skip ↳ posting for the sake of posting, no point of view, no asset ↳ it fills the feed and trains the algorithm on weak engagement The trap is not knowing your ICP enough. If that’s the case, it is very easy to spend creating content (in any format), chasing vanity metrics, and nothing is going to 1) drive inbound pipeline 2) speed up closing due to increased trust and education 3) attract talent, investors, and partners based on your goals Tools to get started: Scripe + HubSpot/ Attio To improve: Apify + Python + Claude So we run a weekly split: mostly MoF to prove it, one ToF for reach, one BoF to convert. Then the engagement on each post feeds the same signal score our SDR is using for outbound, supported by GTM engineers. Content and outbound are not two jobs. They compound when they work together. What share of your posts last month were just reach? Comment SCORE, and I will send you the 4-tier framework we use for both. P.S. We build signal-based GTM engines for B2B teams, where content, ads, and outbound run on one score. If that is the motion you want, book a call through the site or DM me. | 122 comments on LinkedIn

Claude's Fable model finally made my dream of automatically generating leads for companies I've never even spoken with a reality. I didn't lift a finger to launch. I've always thought the best customer experience we could have is to show someone that we can generate leads for them before they pay us anything. So we started running free tests for people to prove it. I've always wished it could happen automatically so it could become a growth motion bigger than what my organic reach / outbound can convert. Fable finally cracked it. From sub agents that source the right fit companies to run campaigns. To building the ICP list in Prospeo, BlitzAPI and GetLeads (depending on the job) To the hardest part, copywriting that would actually convert. Fable was the model we needed to finally handle all of the unexpected nuance of this endeavor. We ran our first 10 completely automatic campaigns and 3 of them got leads. We have multiple steps to never mention the company name that we are testing this for so and to never mention any social proof. We are purely testing if the value proposition resonates with the prospects. We generated leads for Trove RevolutionParts Luxury Presence The replies are all quoted in the graphic below with the ICP and main hook of the messaging. Again, we don't mention any of the company names or anything. Just the value prop so we can validate the messaging at scale. | 28 comments on LinkedIn