Show HN: Find and follow journalists relevant to your startup using ML | Hacker News Reader

Show HN: Find and follow journalists relevant to your startup using ML

app.upbeatpr.com

92 pointsdtran13 comments

Hey everyone,

After helping hundreds of companies with PR and having talked to thousands of founders, the most common thing we hear is: "I'm not sure how to get started with PR" or "I'll think about it in a few months when we're ready to launch." So for hack week, we brainstormed around this idea of "how to get started", and one of the projects we built is a super simple, low commitment and low effort way to get started with PR. Just tell us a little a bit about your company in 30 seconds, and voila, we try to find the most relevant journalists for what you're working on, and let you subscribe to a weekly digest to stay on top of news about your industry, competitors, etc.

Whether you're working at a company, starting a company, or just thinking about starting something, we hope this can help! It's still a very rough v0.1. Feedback, bug reports, etc are would be SUPER APPRECIATED! Thanks so much!

A few different things going into generate the list of 5 journalists that we think are relevant to your company, including a simple machine learned topic model. Happy to give more details if anyone's interested. Some of the results could definitely use some heavy tweaking— feedback is super appreciated!

This is cool, a few ideas:

1) Would like to understand why certain journalists were recommended to me. I like the idea of showing example articles, but the ones provided weren't relevant to me.

2) As a startup, I'm mostly interested in a journalist's influence and reach (ie. how many people are going to see this?). Topic relevance is also important, but at the end of the day I'd rather be written up by TC rather than by a local news org. Does the ML take influence/reach as an input?

3) It'd be useful to select my ideal outlets, e.g. national outlets, industry-specific, geo-specific, and then get results based on those preferences.

Ricky here from team Upbeat. Great feedback. Yeah, the trickiest part was relevance as perceived by users. What our team thought was relevant didn't pass the sniff test in testing. For because a journalist has demonstrated broad interest in AI, but if we only show the article about AI replacing jobs, then she might appear irrelevant to an AI fraud detection company. Sometimes it depends on your familiarity with the publications. Like you said, TechCrunch vs. a more niche B2B publication read by CIOs, we do consider that as part of the input. More customization in search seems like a logical next step.

I really like the idea. Feel there is lots of potential to build similar stuff for anything related to sales.

Hey neerkumar,

Thanks for the kind words! We aren't focused on sales, but in case anyone here is, what do you mean? It'd find sales prospects based on how similar their companies are/what they tweet about?

Finding the right journalists to reach out to so that they can then write about your business can be seen as a special case of sales.

Similarly, I think it would be very useful something like: "I built product X, scrape the web and find me the best sales prospect for it". This is something that people do in a very manual way today and takes forever.

Isn't that what Datanyze and a variety of other products do now?

Thanks for making this service and sharing on HN. Is it possible to provide more contact info for the relevant journalists? I usually rely on Google News alerts to find new & relevant articles for my investment / finance startup. What gets me though is when I find a good article, the author's contact information is usually nowhere to be found so it makes it pretty difficult to do PR.

Hey Amar,

Thanks for checking out our project! What alerts do you have set up in Google News Alerts?

Re: providing contact info— that's definitely one logical next step. We left it out of this hack week project. Journalists already get hundreds of pitches a day, so we're trying to experiment with something new and different in addition to the usual pitch cycles. We're thinking of letting journalists know that you're following them and reading their articles, and make it so they can follow your company as well to receive updates from you, or ask you for quotes/opinions on topics that you're an expert on. This already happens to an extent, but we can make all this much more efficient by only routing their relevant articles to you, and only showing your relevant quotes/updates.

One Google News alert I use is sec filings as that is what my startup offers so I would want to know which journalists are reading sec filings and incorporating sec filing content into their articles. I would then want to contact them to suggest using my startup for sec fillings but your approach makes sense too. Thank you for explaining.

Upbeat is the major innovator in PR today, no one else is even close.

What kind of ML did you use here? How did you deploy it?

Data scientist for Upbeat here. We generated the topics using an unsupervised LDA model trained on a couple million articles from high-quality publications. We took the most intuitive/human-readable of those and used them as the categories you see in the dropdown menu. To get relevant journalists for each company, we scan through our journalist database to find writers whose articles have a high vector similarity to the topics chosen.

Reply on news.ycombinator.com