Seeking Effective Personalization Techniques for Cold Outreach Despite LinkedIn Data Limitations

·May 11, 2026 11:51 AM·

Hi everyone, I’d love to get some ideas on what’s currently working for personalization in cold outreach. I’ve tried a lot of different personalization techniques already, but I’d really appreciate any advice or newer methods that have been working for you recently. One challenge I’m facing is that Clay now has very limited visibility into LinkedIn profiles, since it can no longer reliably pull recommendations or posts older than a week. Would love to hear what kinds of signals, data sources, or approaches you’re using now for meaningful personalization at scale.

🖼️1

9 comments

· Sorted by Oldest

Hisham R.
·
·
what do you mean by techniques? or you mean tools?
Arpit C.
·
·
Hey Rohaan N., it's mandatory for members to add their full names and a profile picture to their Slack profile (pets or AI-generated portraits also work). The community feels more intimate and less anonymous when one knows who they are engaging with so will be great if you can update your profile here. Thanks in advance! 🤝
1
🤝1
Rohaan N.
·
·
Hisham R. I mean, variables/personalisation that we integrate within cold outreach
Viky G.
·
·
I tried an interesting one which is not so common - means I’m not sure if there is a variable for it - but I checked if the potential lead writes about the pain points that are related to the product I work with. And then I crafted my messaging based on those. Keep in mind they still need to be reviewed closely, especially if it’s a technical product, because there may be a lot of relevant information found by AI. That said, a lot depends on your ICP - e.g. for folks that we had a ton of out each messaging, you have to be really witty to get their attention
Hisham R.
·
·
keep your first message around 75 words no links no images, try to use claud for intent search and find a pain point to personalize around it, i hope that help
Rohaan N.
·
·
Thanks Viky G. Which platform you were scraping through to find out if they have mentioned the pain point, was it linkedin? I have trouble ClayAI Scrapers (Claude, Gemini, GPT) scrolling through the person's linkedin profile and see their posts/recommendations and these sort of things
Jesse O.
·
·
Hey Rohaan N. Implementing a job-change reactivation workflow might help and no manual LinkedIn investigating. The triggers let you know about any new roles, new companies, and the date of any changes. Here's a course on how to implement with that with Lusha.
Jack S.
·
(edited)·
Rohaan N. We're a B2B IT company and our signals lean away from LinkedIn and more into timing and intent. Website visitor identification is the big one. If someone's on our site reading specific service pages, we enrich the company, find the decision maker, and tailor the message based on what they looked at. Beyond that, job changes into IT decision-making roles, job adverts that suggest scaling or tech gaps, and news signals like funding, new offices, or acquisitions. Anything that suggests infrastructure decisions are being made. Basically personalising around what's happening in the business right now rather than what someone posted online months ago. We're still developing ideas but that's the direction.
Stefan D.
·
·
Rohaan N. Actors are underrated here. If you're scraping LinkedIn or any paginated source, they handle scale way better than standard scrapers and help close the data gaps you're running into. On personalization: tokenize your highest-leverage variables. Figure out the 2-3 signals that actually move the needle for your ICP and build around those. Everything else is noise.

Hisham R.
·
·
what do you mean by techniques? or you mean tools?
Arpit C.
·
·
Hey Rohaan N., it's mandatory for members to add their full names and a profile picture to their Slack profile (pets or AI-generated portraits also work). The community feels more intimate and less anonymous when one knows who they are engaging with so will be great if you can update your profile here. Thanks in advance! 🤝
1
🤝1
Rohaan N.
·
·
Hisham R. I mean, variables/personalisation that we integrate within cold outreach
Viky G.
·
·
I tried an interesting one which is not so common - means I’m not sure if there is a variable for it - but I checked if the potential lead writes about the pain points that are related to the product I work with. And then I crafted my messaging based on those. Keep in mind they still need to be reviewed closely, especially if it’s a technical product, because there may be a lot of relevant information found by AI. That said, a lot depends on your ICP - e.g. for folks that we had a ton of out each messaging, you have to be really witty to get their attention
Hisham R.
·
·
keep your first message around 75 words no links no images, try to use claud for intent search and find a pain point to personalize around it, i hope that help
Rohaan N.
·
·
Thanks Viky G. Which platform you were scraping through to find out if they have mentioned the pain point, was it linkedin? I have trouble ClayAI Scrapers (Claude, Gemini, GPT) scrolling through the person's linkedin profile and see their posts/recommendations and these sort of things
Jesse O.
·
·
Hey Rohaan N. Implementing a job-change reactivation workflow might help and no manual LinkedIn investigating. The triggers let you know about any new roles, new companies, and the date of any changes. Here's a course on how to implement with that with Lusha.
Jack S.
·
(edited)·
Rohaan N. We're a B2B IT company and our signals lean away from LinkedIn and more into timing and intent. Website visitor identification is the big one. If someone's on our site reading specific service pages, we enrich the company, find the decision maker, and tailor the message based on what they looked at. Beyond that, job changes into IT decision-making roles, job adverts that suggest scaling or tech gaps, and news signals like funding, new offices, or acquisitions. Anything that suggests infrastructure decisions are being made. Basically personalising around what's happening in the business right now rather than what someone posted online months ago. We're still developing ideas but that's the direction.
Stefan D.
·
·
Rohaan N. Actors are underrated here. If you're scraping LinkedIn or any paginated source, they handle scale way better than standard scrapers and help close the data gaps you're running into. On personalization: tokenize your highest-leverage variables. Figure out the 2-3 signals that actually move the needle for your ICP and build around those. Everything else is noise.