SDRLabs AI

Another great update! Want to learn more - check this out: aisdr.com/?ref=sherman71

Going into 2025 we are excited for our strategic partnership with the team at AISDR . Keep up all the great work team!

So what’s better: X/Twitter or LinkedIn? I’m not a social media expert, but we will end the year with about the same views (40,000,000) and followers (250,000-300,000) on both So at least I can share a few learnings: #1. X and LinkedIn perform about the same in terms of driving traffic It used to be different, but in 2024, they both drive about the same amount of traffic to SaaStr.com (about 15% total) and about the same amount of attendees to events #2. Yes, you can re-use content across both, and they perform differently The audiences are distinct enough that there appears to be little downside to cross-posting to X and LinkedIn. Posts that take off on one often do not take off on the other. But a post that is not popular on one rarely is popular on the other. #3. Time of Day doesn’t seem to really matter Near as I can tell, it doesn’t really matter when you post to X or LinkedIn, at least not that much. Either a subset of your base quickly engages or they don’t, and the algo pushes it out. Now for our newsletters, where there is no algo push, time and day do make a huge difference. Evenings and Fridays way underperform there. #4. X and LinkedIn are both equally as terrible is supporting links to your own site, either in the comments or the post itself The best practice of putting the link in the comments on LI, or a second tweet on X, does work in the sense that you aren’t penalized for doing so. But we get almost no traffic from those “link in comments”. They are barely worth it at all. Both platforms are now terrible for driving traffic to your own sites. It is what it is. #5. Company pages on LinkedIn perform terribly A lot of folks have written this, but we have 65k on our company page and 280k that follow me — a 4x difference. But the engagement is 100x higher on my Jason LI than SaaStr LI page. 100x. #6. Adding an image, short vs. long posts, “hooks”, etc., none of it seems to matter Or if it does matter, it doesn’t seem to matter predictably enough. #7. For B2B at least, the conversations are 10x better on LinkedIn I think this is likely a combination of both audiences and how the platforms encourage comments and debate. LinkedIn is a much better place to engage with B2B conversations with B2B professionals. But … #8. CEO-level engagement is much, much higher on X than LinkedIn At least for me/us, LinkedIn does not seem to be anywhere near as strong a place to engage with CEOs and founders. #9. You really do need to own your own list, somehow LinkedIn, X, YouTube, Spotify, etc. are all great. We now have to invest much more in all of them than just a few years ago. But they also are all in the end terrible for driving folks to your own sites and platform. They are all walled gardens. I don’t think a crappy list has much value at all. But a list of 10,000+ superfans that you can email directly? Gold.

SIX major updates in 21 days.... The team at AISDR have been busy - here is what's been shipped in December so far: 1. Website visitor identification and outreach. De-anonymize your website visitors and convert them. Multiple providers for maximum coverage; instant filtering on ICP; fast, personalized outreach via email and LinkedIn. 2. 10x deeper person level research. We used to offer only LinkedIn profile and posts for person-level research. Now, we offer Youtube, podcasts, X (twitter), interviews, personality traits and of course, LinkedIn activity. 3. New data vendor. It's important for us to allow our customers to pick the data vendor with the most coverage for their audience. Many rely on LinkedIn Sales Navigator, so now we've added the data vendor that mimics its search structure as close as possible. 4. Campaign scheduling. Sometimes, especially during holidays, you want to set everything up, but start the actual outreach later. Now you can easily schedule campaigns in AiSDR to start sending after the holidays or anytime. 5. Campaign page redesign. Because the product needs to be beautiful and easy-to-use and not only useful. 6. Automated email credit management. We have automated email credit carry over to next periods. If for some reason you haven't used them within a month - you will have more credits to use next month. Want to learn how to get started with #AISDR and get more meetings booked? Book a call with SDRLabs AI here: https://bit.ly/49UkoLt

It's Day 19 of the 24 days of the Go-to-Network Advent Calendar! We teamed up with Commsor 🦕 to bring you 3 ways to automate outreach without losing a human touch. 📅 Open today's gift and get the plays from their virtual advent calendar 👉 commsor.com/advent 🎁 Subscribe to Warm Revenue Weekly 🔥 to have the chance to win: A 6-month free pilot of Close (subscription only, usage not included), with a white-glove onboarding experience and a sales process consultation with us! Subscribe here if you haven't already 👉 commsor.com/newsletter Happy selling!

ChatGPT can now work directly with more coding and notetaking apps—through voice or text—on macOS. • Work with your code in context with expanded support for coding apps like Warp, IntelliJ IDEA, PyCharm, and more • Code with the power of o1 and o1 pro mode • We’ve added support for note-taking apps like Apple Notes, Notion, and Quip. • And you can use Advanced Voice when you work with these apps. This beta is available to Plus, Pro, Team, Enterprise, and Edu users. Download the ChatGPT app for macOS to try it out: https://lnkd.in/eeChPQTf We plan to bring this feature to Windows and Free users next year.

It's time to spread the holiday cheer 🎊 Follow your friends on Phantom and share your profile to social media. You might just get a follow back from members of the community… or us 👀

If you're using Apollo and you're not cleaning your data, you are shooting yourself in the foot and killing your campaigns. Here's why: The Test Results We tested a 200-sample lead list in SaaS and found: - Apollo provided 125 emails (62.5%). - After validation, only 74 were valid (37%). Why Is This Bad? - Hurts domain reputation: High bounces lead to spam flags. - Lost leads: Invalid emails = fewer opportunities. - Wasted money: Rebuilding email infrastructure adds overhead costs. What to Do About It? Use Persana. With 700M+ contacts and access to 75 data providers, including Findymail, Icypeas, Leadmagic, Prospeo, and Datagma, you can perform a waterfall enrichment to get: - Better coverage. - Cross-verified accuracy. - High-quality, actionable data. Results When we enriched the same 200 leads using Persana: We found 158 emails, with 121 valid emails (a 37% improvement over Apollo). So what are you waiting for? Try Persana today and experience the difference with cleaner, more accurate email lists. ✅ 14-day free trial. ✅ No credit card required. Link in comments

New feature 🚨 Campaign scheduling 📅 Now you can plan ahead and set the perfect launch date for your campaigns 📧 It’s as easy as scheduling an email. No rushing. No forgetting. Just smooth, timely outreach Book a demo with us to see how SDRLabs AI can get you up and running with AISDR

Claude 3.5 Sonnet vs. GPT-4o vs. GPT-o1 for sales emails - my 4-hour prompt experiment. So today I spent a couple hours prompting. In the past year and a half I got really good at it, having created and iterated on thousands of prompts with AiSDR. My goal today was to test out new prompt structures with our new enriched person-level data and latest models. TL;DR - Claude 3.5 Sonnet is the best performing model for emails; - o1 is second best, but very expensive and unstable (request fail often via API); - 4o is 3rd, but there is an interesting nuance. Emails are attached. I used my co-founder (less active on social media) and I (more active) as prospects and a random company for value prop (we are also not their target audience, which was part of the challenge). Each email was created using the same context and the same prompt. The reason Claude performed better is because it's overall writing capabilities with "loose" prompts are much higher. If you are not a pro in prompting - Claude is your absolute go-to-LLM. 4o writing has all the typical "AI email" red flags: "noticed, impressive, etc.". The only way to get rid of those and get 4o to write great emails is with very prescriptive, structured prompting. With prescriptive, structured prompting Claude actually performs worse, than 4o. It feels like Claude was created to "write freely". o1 is in the middle. I think it is actually close to Claude on writing, but because of the cost and because it has much stricter limitations on calls via API, it's not a good choice for this task. Would love to hear your experience and thoughts on these emails.