Expecting top tier coverage from your user conference? You’ll need a little help from your friends

I hate to be the one to tell you, but Hot IPO Summer is almost over, and that time-honored tech tradition, User Conference Autumn, will be here before you know it. VMWare Explore, the unofficial opening ceremony, kicks off in Las Vegas at the end of this month. In just five weeks, 40,000 lanyard wearers will file into Moscone Center for Dreamforce. Break out the hand sanitizer!

So much time, money and effort go into a good user conference, ratcheting up expectations for all facets of the business. Sales has well-defined targets, Product has their roadmap, CS needs to book in trainings… but ROI for Comms can be murky. Too often, we see “keep the executive spokespeople busy” and “get more earned media than last year” as goals. Big, conversation-starting coverage is elusive, even among household names like Salesforce and AWS. But surely quality is more important than quantity? How do boards, investors, and potential customers move based on news from the user conference?

The method

To better quantify the opportunity for top tier press, we activated our in-house AI research function to analyze 50 B2B technology user conferences — one instance of each, all held between July 2025 and June 2026, every event owned by and built around a single brand. So Dreamforce, re:Invent, and your various Connects and Summits are here, but no massive industry events like Mobile World Congress or TechCrunch Disrupt. We tracked which publications covered each event, in a time window running from a week before the doors opened (scene setter coverage) to two weeks after they closed (wrap-ups).

Then we counted how many reached the upper echelon of tech press. In this case, broadcast or online coverage from a representative sample of CNBC, Bloomberg, TechCrunch or Axios.

Tier One is tough, but solvable

Out of our group of fifty, only fourteen garnered top tier coverage. Four of the fourteen were hyperscalers: AWS, Microsoft, Google and Oracle. Three more were megavendors: NVIDIA, Salesforce and Adobe. The remaining seven were HPE, ServiceNow, Zscaler, Databricks, Alteryx, GitHub and OpenAI.

Three of the thirteen large public companies in our panel got there: HPE, ServiceNow and Zscaler. Not one of the eleven mid-caps did. And of the three large-cap hosts that broke through, only HPE managed it through reported coverage of something it announced.

If you run a flagship comms program at a multi-billion-dollar software company, and have been disappointed by the business press coverage coming out of your company’s annual user jamboree, that mid-cap zero is probably the most useful number in the study. It says more about how editorial desks decide what counts as news than it does about anyone’s execution.

However, the fourteen who broke through did something specific, and it’s repeatable once you’ve seen the pattern.

Three were too big to ignore

AWS, Microsoft and Google Cloud are big enough that their own roadmap was the story. Great for them, but nobody else in the study had that option.

Seven brought in a credible wingman

  • Oracle earned CNBC with a headline that read “Oracle Cloud to deploy 50,000 AMD AI chips, signaling new Nvidia competition.” The story was not about Oracle’s core operation at all, but instead about their bet on AMD in the face of price and supply pressures on Nvidia.
  • GitHub earned CNBC for Agent HQ. Everyone announced in-platform AI initiatives, of course. But this one is notable because they brought OpenAI, Google and Anthropic agents together in one control plane.
  • Salesforce earned TechCrunch for Agentforce 360, where the coverage led on an expanded OpenAI partnership and support for Google’s Gemini.
  • OpenAI earned CNBC and TechCrunch for the Apps SDK, demonstrated through Zillow, Canva and Spotify.
  • NVIDIA earned Bloomberg and CNBC for Vera Rubin — and note who carried the story. OpenAI, Anthropic and Meta were not partners on that platform, they were buyers. The news was not the new hardware architecture itself, but instead the big-name customers committing to big orders.
  • HPE earned Bloomberg for Discover, under the headline “HPE Boosts Networking Gear for AI Clients, Adds Siemens Energy.” A named customer, doing something specific with the technology, in the headline.
  • Adobe put Shantanu Narayen on stage with NVIDIA’s Jensen Huang, and CNBC covered the conversation. With its own roadmap under question, Adobe borrowed someone else’s credibility.

A product launch says something about your company; a partnership with headline players tells a story to the market. Once a second organization has publicly committed money, engineering or distribution, the announcement carries information about where the category is heading, and that’s what a Tier One business desk is interested in.

It also clears the “third party validation” hurdle. Editors heavily discount what a company says about its own products, and reasonably so. They discount it much less when AMD or Zillow has put its own engineering time behind the claim.

Two private companies shared great revenue numbers

As private companies, Databricks and Alteryx don’t HAVE to publicly announce results. If they want to tell the financial press how the business is going, they can decide when and how.

Both used their own event as a disclosure vehicle. Databricks put annualized revenue growth of more than 80%, at $6.9bn, center stage at their Data +AI Summit, which provided a superb setting for an interview with CNBC.

Likewise, Alteryx’s Andy MacMillan went on CNBC with $1bn in ARR — a milestone that matters mainly because the company is being readied to go public again.

They are two great examples of the comms and events teams working together, taking a piece of good news and turning it into something memorable for the employees and customers in attendance, while raising the flag for the analysts, investors and prospects watching at home. A private company has no earnings call, no guidance, no quarterly drumbeat, so their user event can do a lot more heavy lifting than their publicly traded counterparts.

Two public companies saw an opportunity to clear the air

ServiceNow and Zscaler both reported earnings in the waning days of Spring’s SaaSpocalypse, and courageously used their user conference to set the record straight. Zscaler’s shares fell 31% on results on 27 May, and Jay Chaudhry was on CNBC twelve days later, from the floor of Zenith Live. ServiceNow’s price fell 18% on 23 April, and Bill McDermott was on CNBC twelve days after that.

The inconvenient truth is, a difficult quarter can be the thing that makes your CEO bookable. Most comms teams give in to the urge to keep the CEO away from cameras, in exactly the quarter when the press would take the call.

For ServiceNow and Zscaler, lining up the post-earnings CEO interview at their own user conferences was a masterstroke in framing. Their customers in the room, and a reporter who suddenly has a reason to turn up, gave McDermott and Chaudhry a home field advantage, with loads of announcements fresh from the show floor to turn the page. I’m sure it took a lot of trust from the CEO, but the resulting interviews became an asset for sales and IR for the rest of the summer.

The bulk of meaningful coverage is not in Tier One

One correction worth making while we are here. Most conference media plans are built around a Tier One wish list. The publications that reliably turn up are somewhere else entirely.

Across our panel, SiliconANGLE covered 30 of the 50 conferences. Futurum covered 23, TechTarget 20 and diginomica 16. Combined, CNBC, TechCrunch, Bloomberg and Axios covered just 14 of the 50.

All 50 conferences, and who covered them

Each row is one conference, grouped by company size and ranked by outlets reached. Each column is one publication. Filled cells indicate coverage in the window. Rows scoring 9 or 10 hit the limit of our search method, so their true breadth is higher.

 

These trades are analyst-adjacent media, and they’ll happily spend time and column inches explaining enterprise technology to the people who buy it. Most importantly, their audience is highly concentrated with potential buyers and decision makers.

Partners help here too, for the same reason: a named integration gives a trade reporter a concrete thing to assess.

Applying the pattern

Build the conference around one primary announcement with somebody else’s name on it. A partner with something genuinely at stake, quotable, and willing to appear in the release — a customer logo slide won’t do it. If you can’t name that partner eight weeks out, the keynote won’t rescue it.

Influence the influencers. Our prior research on the indie journalism movement came in handy here. In the last eighteen months, we’ve seen a wealth of credible, respected tech journalists leave their established mastheads for a solo publication hosted on Substack or beehiiv. These reporters often have seen a lot, carry a tremendous amount of trust with their audience, and can work with brands in new ways that their legacy outlet would have forbidden. Starting with a flight to your user conference.

Decide what the single, cohesive story is. Volume of announcements can work against you. Our research explored a year where development teams were spoiled with an abundance of AI developments. Everyone announced AI news, but a few host brands piled on feature announcements, ultimately garnering much less coverage than their peers. If the news can’t be reduced to one sentence containing a number or a second company, you have a product update, and product updates live exclusively in the trades.

A closing note on what we didn’t measure. We tracked earned media, and separately checked share price, where across 38 public hosts we found no detectable conference effect at all. Of course, we can’t see into private pipeline, renewals, customer sentiment, or the value of getting several thousand customers into one room, which is, in all likelihood, why these events get funded in the first place. It just won’t get you on CNBC on its own.

Methodology note: 50 host-run B2B technology user conferences, one edition each, held July 2025 to June 2026. Coverage measured across a fixed 42-publication panel, in a window running from seven days before each event to fourteen days after. Three events are hosted by brands owned by larger companies (GitHub Universe, Red Hat Summit and Splunk .conf) and are classified by the brand running the event rather than by its parent.