I Sent 41 Personalized Press Releases in One Afternoon
Personalization at scale is the difference between spam and coverage. Here is how we built a system that treats every journalist like the only journalist.
Most press releases die the same way. Someone writes one version, blasts it to a list of 50 journalists, and then wonders why nobody replied. The release hits the inbox looking exactly like everything else in there -- generic opener, boilerplate body, no indication that the sender has any idea who they are emailing.
Journalists delete those in seconds. If you have ever watched engagement on a generic send, you know.
What actually works is personalization. Not fake personalization -- not "Hi [FIRST NAME]" -- but an opening paragraph that demonstrates you read their last three bylines, you know their beat, and you are pitching them something that fits. That is the gate. Everything after that is just good writing.
The problem with real personalization is time. Writing 41 custom openers by hand takes most of a day. So we built a system to do it in about two hours of prep and one afternoon of execution.
How the System Works
We maintain a journalist contact database with four fields beyond name and email: their publication, their beat (two or three words -- "energy markets," "Southeast Asia tech," "fintech regulation"), their three most recent article titles, and a one-line note on what angle they respond to.
That last field is the important one. Some journalists love data. Some want a narrative. Some only care if there is a Bangkok angle. You learn this over time and you store it.
When a press release goes out, our system uses those fields to generate a personalized opening paragraph for each recipient. Not the full release -- just the first 60 to 80 words. The rest of the release is the same for everyone. The opener is unique.
For an energy desk reporter, the opener referenced a recent beat piece and framed the announcement through that lens. For a Bangkok tech desk, it opened with the Thailand operations angle. For a wire reporter, it led with the single data point most likely to be quoted in a headline.
The Technical Setup
I use SendGrid for delivery. Not because it is the cheapest -- it is not -- but because deliverability matters more than cost per send when you are trying to reach major business-desk inboxes. You get one shot. If the email lands in spam the first time, you are done with that contact for months.
Every email goes out with both plain text and HTML versions. This is non-negotiable. Some journalists read plain text. Some email clients strip HTML. If you only send HTML, you are gambling. Send both.
The HTML template is branded -- it has the company logo, a clean single-column layout, and the release formatted like a real document, not a marketing email. No stock photo headers. No gradient backgrounds. It looks like something a communications team sent, because it should.
Sends are staggered. Not random staggering -- scheduled staggering. Global wire and business desks go out at 8:30 AM their local time. Regional Bangkok titles go out at 9:00 AM Bangkok time. Nobody gets an email at 3:47 AM their time because my server sent the batch all at once.
What the Sends Looked Like
Forty-one contacts. Six publications. Three brands, each with its own release and its own SendGrid sender identity -- because QFHQ announcements do not come from the same sender address as OilMarketCap announcements, and they should not. The statement descriptor, the from name, the reply-to -- all matched to the brand being announced.
Total prep time: about two hours building the contact segments and reviewing the personalized openers before sending. Actual send time: under ten minutes once the queue was loaded.
Early engagement on personalized sends has consistently outperformed generic blasts in my experience. Specific open-rate numbers are campaign-dependent and not something I'll post without context — but the qualitative gap, measured by replies and follow-up questions, is large. The difference is entirely in whether the recipient feels like the email was written for them.
What Most People Get Wrong
They treat press outreach like email marketing. It is not. Email marketing is volume. Press outreach is precision. You are not trying to convert a percentage of a list -- you are trying to get specific humans to make a judgment call in your favor. That requires a different mindset and a different system.
The other thing people get wrong: they send the same release to reporters who cover completely different beats and then wonder why only some reply. A tech reporter does not care about your operational infrastructure unless you frame it as a technology story. An energy reporter does not care about your AI unless you frame it as a market impact story. You are not changing the substance -- you are changing the angle. And you have to do that per journalist, not per publication.
The Takeaway
Forty-one sends. Two hours of setup. Plain text and HTML both. Staggered by timezone. Personalized openers per beat. Each brand sending from its own identity.
That is not complicated. It is just disciplined. Most people skip the discipline because it takes more time upfront. But the return on that upfront investment -- in actual coverage, in reporter relationships, in brand credibility -- is not comparable to blasting a generic release and hoping someone bites.
Build the system once. It runs every time after that.