Sangeetha Arun’s Post

Your inaccurate signals are ruining any chances of an email response. BUT - if you have the right data you can make some pretty cool emails. Observability people will get this message. Check out how we looped in observability signals into this email 👇 It’s just a framework - I know a lot of folks love an interest based CTA! But shows how having the right data can help craft a much more relevant message. Hit me up if you want to see if our platform can get the data you need!

Excited to share my 4th project - Email Spam Detection model In the ever-evolving landscape of digital communication, accurate spam detection is crucial for individuals and organizations alike. Our project aims to revolutionize this process, harnessing the power of data science to identify and flag spam emails with precision and reliability. Project Overview: Our journey into Email Spam Detection has been a fascinating one, driven by our passion for data-driven insights. By analyzing a wide range of email features and leveraging advanced modeling techniques, we've developed a robust predictive tool that empowers users to distinguish between legitimate and spam emails with confidence. #EmailSpamDetection #DataScience #OasisInfobyte

🚨 New Enrichment Alert: the Email Format & Alexa Rank enrichment tool verifies the structure of email addresses in extensive lists, sorts various categories of email addresses, and supplies the Alexa traffic position for the domain. You're welcome 🤗 https://buff.ly/49txj6b #bigdata #nocode #spreadsheets #dataanalysis #dataops

Email Spam Detection Model I'm pleased to share my latest project focusing on Email Spam Detection using Logistic Regression. This model accurately predicts whether an email will be categorized as spam or not, making it valuable for optimizing email marketing campaigns. Project Highlights Algorithm Used: Logistic Regression Accuracy: Achieved 96% accuracy on both training and test datasets Application: Designed to help businesses filter out spam emails effectively, improving email campaign efficiency and user engagement. This project showcases my dedication to developing practical machine learning solutions with tangible business benefits. #MachineLearning #DataScience #PredictiveAnalytics #LogisticRegression #EmailMarketing #MLModeling

turns out it's important to get your pitch right! This is especially hard if you're an engineer. It's really easy to get caught up in pitching the product as a list of features, rather than the problem it solves. For us, it was "unstructured data". Which seemed pretty clear to me! Things like documents, text, websites, audio files, etc. But when we pitched the product, we mostly got: "oh interesting, but we don't have any unstructured data." Then two weeks ago, we changed the pitch to "We read your PDFs". And overnight the traction skyrocketed. All with no change at all in the underlying product. Plus the quality of inbound leads went way up. People had exactly one problem that needed to be solved. Including customers we talked to before! "We don't have any unstructured data, but we have tons of PDFs!"

Data does not lie. This context: Email Deliverability . There are countless tools that will tell you if your email is fine or not. Some good some not. But here's a data analysis approach that can be used to discover such trends, problems. With 100% accuracy. So the problem was related to recent Gmail/Yahoo changes. Our methodology was simple: - analyze user behaviour historically in the past 30/60/90 days - differentiate between gmail and non-gmail users - results interpretation. You just need a very capable CDP that collects this data right. Obviously this is a very simple example, but is a good start #cdp #deliverability

If something like the subject line can make or break your emails, There's only 1 choice. Do everything you can, Split testing, optimization, emojis, text, case (lower case/sentence case), EVERYTHING. to ensure it makes, Not breaks your emails. and break it will, someday. When it does, fret not. Coz you've got a lot of data to analyse and put your best foot forward.

Think adding {{firstname}} in your subject line is real personalization? I’ve got news for you—it’s not. And it won’t magically boost your open rates either. Lemlist analyzed millions of emails, and the data shows adding “you” or a first name made no significant impact. Open rates stayed around 34-36%, with or without it. No tricks here. The real key? Personalizing based on context, behavior, and timing. It’s about solving problems and tapping into specific actions—not just throwing someone’s name into a template. So, forget the “Hi {{Firstname}},” and start thinking about what's important or contextual to that person in the moment. Because basic personalization doesn’t drive action—relevance does.

Cold Email Tech Stacks Are Expensive! But there are ways to reduce costs without compromising quality. Here’s the tech stack for sending 1,000-2,000 emails/day: 1. Scraping Companies from Niche Directories: → SAAS: Crunchbase → Agency: Agency Vista → Ecommerce: Store Leads → Local Business: Google Maps Estimated cost: $0-$100/month 2. Email Enrichment + Verification: → Apollo.io: $50/month → Findymail: $100/month → Million A.: $50/month 3. Data Point Enrichment using Persana and OpenAI: → Persana AI: $85/month → OpenAI credits: $50-$100/month 4. Email Sending Tool: → Smartlead: $100/month Estimated cost for this tech stack: $600/month (excluding inbox and domain costs) This is almost 50% cheaper than the average tech stack cost for sending 1,000-2,000 emails/day, which ranges from $1,000-$2,000. PS: What are you spending on your current tech stack?

Our brand new tech blog is live! 🙌 First up: Senior Data Scientist Rok Piltaver delivers a decade’s worth of A/B testing wisdom, including some hard-won lessons learned. Find out how data-driven decisions can enhance user experience, boost monetization, drive business growth, and more in the first instalment of this two-part series. 👉 https://lnkd.in/gqDFxz9e