STH Q2 2024 Letter from the Editor Change and Limits

Patrick Marvell Accelerated Infrastructure AI Day 2024 Large
Patrick Marvell Accelerated Infrastructure AI Day 2024 Large

Every quarter, I like to do a little update to give our readers a behind-the-scenes look at what is happening. Often, there is a big difference between what folks see publicly and the inner workings of STH, so I like to peel that back. In this edition, I wanted to share our Q2 2024 update and really get into some of the big changes we need to make, the industry context, and why. I also have an ask of our readers.

Previous Updates

If you want to check out how this series has evolved, here are the links to the previous ones:

We now have been doing this for over five years. It is amazing how time flies.

2024 Q2 The Realization Hit Me

Looking back on the STH Q4 2022 Letter from the Editor, it was almost prophetic, but it was missing a key component: the AI side. The publishing industry is changing at a rapid rate, and your favorite tech sites are under pressure. This is going to force STH to make a bit of a “lane change,” which we have been foreshadowing for many quarters already. I want to explain why, after 15 years of STH, this must happen.

Today, web publishing is in terrible shape. While the STH main site is growing, it is no longer doing so at a 23-25% CAGR. The trend is clear. With companies like Google and Microsoft placing AI-generated snippets at the top of search results, organic results fall another place further down the page. One of the reasons folks point to the threat of AI to Google’s dominance is that Google dominated the search and discovery process for a long time. Essentially, the web is now built around Google’s search and discovery. Generative AI, even by disrupting only 10% or 25% of what is being done today, is going to kill off publishers. I sense that the impact will be bigger.

Proxmox VE Backup Change Time
Proxmox VE Backup Change Time

For publishing outfits, this will have significant implications. If someone asks how to do something, those questions will be answered without going to a website. Learning what the best is, keeping up on the news, and so forth is going from a Google search or a news app today to a generative AI or voice assistant soon.

What has happened in the publishing space has been rough. What might surprise folks is that STH has gone 15 years without invoicing companies directly. Companies support STH by buying ads through third-party agencies, and that is what gives us the budget to do projects. We discuss projects we want to do with companies, but we have been lucky. I can say, “This is what we want to do,” and folks help make it happen.

Kioxia CM7 PCIe Gen5 NVMe SSDs In Two Broadcom External DAS Chassis Cover
Kioxia CM7 PCIe Gen5 NVMe SSDs In Two Broadcom External DAS Chassis Cover

Over the years, STH has become huge. We are by far the #1 server, storage, and networking review site in size. The number #2 site is around 1/8th-1/10th of our size in the data center gear review space. When it comes to video, we are in a middle ground now between many IT influencers and massive channels like LTT.

For some sense of scale, we have averaged well over 100K views/ video over the past year. Our target on daily main site articles is also 100K+ views/ article. When we create an article and video, the content is likely to get over 250K views. On the consumer side, these are not huge numbers. On the enterprise/ data center side, they are tremendous. For some sense of scale, NVIDIA is the hot company right now and puts a lot of money into its blog, making it one of the best in the industry. STH usually gets more monthly page views on NVIDIA content than NVIDIA gets on its blog. To put this into another perspective, many of the “Analyst-Influencers” we discussed in our Q4 2022 letter are lucky to get 250 organic views on pieces they create.

Patrick With NVIDIA DGX GB200 NVL72 At GTC 2024
Patrick With NVIDIA DGX GB200 NVL72 At GTC 2024

Naively, I thought that being considered “press” and hitting crazy metrics in our segment would lead to more opportunities to do the content I wanted. Naturally, organizations want to optimize getting the most eyeballs on their content, right? Time and time again, this hypothesis has been incorrect. Companies optimize for analysts and influencers over the press because they can more tightly control the messaging, even if it means reaching 1/1000th of the audience. It might seem strange, but that is how the world works.

Smaller organizations like STH face a challenge in the ad space because we have a single property. Large chipmakers, OEMs, and so forth require a minimum ad buy with a publisher of $2-10M. Publishers like Future that run sites like Tom’s Hardware can have those conversations, but those companies cannot go to STH directly because we are too small.

AMD EPYC 9754 Bergamo AMD EPYC 9684X Genoa X AMD EPYC 9654 Genoa 1
AMD EPYC 9754 Bergamo AMD EPYC 9684X Genoa X AMD EPYC 9654 Genoa 1

Our option would be to lean heavily into ad networks. If you have been to many well-known tech sites, they have ads on the top, bottom, middle, sides, auto-play ads, and so forth, with 300+ cookies being placed on a page view. At STH, we have almost eliminated all ad networks, especially on the desktop. On the desktop site, we eliminated almost all of the “fill” for unused inventory. I have told the agencies buying and selling ad space on STH that we will only accept static banners. Just to give you some sense, even converting our inventory to nowhere near the aggressive levels at big publishers like Future and going through agencies, we are giving $25K/ month at the low end in ad revenue. On the other hand, I hate going to sites with aggressive ads, so I try to keep STH as something I would like to visit.

For us, that is a big challenge. We have crazy amounts of traffic (for the enterprise space), but we limit how we monetize it. Some smaller storage and server review sites will allow vendors to publish articles on their sites or have back-end testing agreements for hundreds of thousands of dollars per year per vendor. I have a pretty firm belief that if I go to a site, I want to know what those people think, not what vendors write about themselves. As such, that is why I say “no” when people ask to pay for that kind of content. There is a dynamic and well-known data center-focused site with an editorial team that is the size of STH. They publish more in a day than we do in a week. They generate tens of millions of dollars in revenue but half of our page views because the line is poorly defined.

Between my stance on ads and my stance that all content on STH must be written and reviewed by our team, many in the industry think I am silly. They have a point.

Over the past year or so, as we have ramped up our YouTube channel, the costs of creating content have increased. It is one thing when I need to travel solo. It is another when we also have to have 1-2 camera people, gear, dedicated space for studio sets, and so forth. The feedback has been, “Why not just be an influencer?”

On its face, being an “influencer” is probably a good fit for STH in today’s world, but I hate the label almost as much as I am averse to calling STH “press.” I am not fond of competing with generative AI posting news pieces. Interviews in the industry tend to be abysmally low traffic unless the interviewer and/ or interviewee are incredibly noteworthy. Basically, due to my preferences and shortcomings, STH sucks at being “press.”

About 18 months ago, something strange happened. More companies started looking at STH as an “analyst” or “influencer” precisely because we do not do “press” things well. As I started talking to folks on that side, I realized this might be the best way to fix the cost issue. For example, by having the Axautik Group as our “analyst” firm, we can monetize our work directly. What has freaked people out is that I have been giving very simple guidance in exploring these paths:

Organizations can help provide resources to make content but not pay to insert their thoughts.

That may seem simple, but it freaks people out. Not giving editorial control, not doing script reviews, not doing final work product reviews, and so forth is almost unheard of in the industry. Everyone expects control. If you want to know why Ubiquiti does not send products to STH, I refused to sign the NDA that gave Ubiquiti the right to see content before it is published and potentially prevent publication. I was told that even very large YouTubers signed the agreement. My response was “no” and has continued to be “no” since that violates the principle of any sponsorship in material or financial being for the time and costs of content creation versus being able to control our thoughts.

Ubiquiti USW Leaf Switch Front
Ubiquiti USW Leaf Switch Front

Saying “no” has been liberating. I realized that STH is big enough that alternatives are maybe 1/10th as effective. Perhaps that has helped us since we have an established track record, and I can say “no.” These days, I am just pitching what I want to do in Q3 and beyond. Companies can support it if they want.

Transitioning from tech press to something else is never easy, but it must happen if we want to keep creating top-bin content on STH without having the site overridden by ads, running only vendor-written or vendor-approved content, and so forth. Part of having the Axautik Group is that it gives us an immediate reason to be considered an analyst firm and still get the same (if not better) access than we would have as “press.”

My goal about a year ago was that we could maintain the status quo through the transition. Something else happened entirely. Now that we are telling companies we no longer want to be “press” since we have not done “press” things for years, the opportunities are much better. For example, we are going to be doing something at the end of July and publishing in early August, in a place and at a cost to film that there is no chance we would have been able to get access to as “press” supported only by ad revenue. We would not have been able to even get in the door as “press.” We have another one in earlier stages that would be doing something insanely cool for publishing in September. Part of my hard line has been we need things to get better with the transition.

HPE Model Server At GTC 2024
HPE Model Server At GTC 2024

We have been stuck where it is hard to do things in our ad-supported model beyond doing reviews in the studio. Even if companies paid for flights, we were out of our production budget, going even slightly beyond a standard server review. That led to us having to cut the budget on other pieces to make up the difference each month. Frankly, we are stuck, and it gets very tedious when you are growth-minded.

TRENDnet TEG 3102WS Dual SFP Plus
TRENDnet TEG 3102WS Dual SFP Plus

This is a lot of rambling, but it has a point. Being “press” sucks and is only getting worse. The programmatic ad-supported model leads to a terrible user experience on the site and one that I do not want to push on others. Going analyst-influencer five years ago for STH would have led to giving up on a fundamental tenet of creating content with independent thought behind it. At this point, making a label change we can say we want to make the same content, but get help to create new and better content.

I know some folks will lose themselves with this, but it has to happen because the traditional press and ad-supported model is dying. We have to make the transition now when I can dictate how the transition occurs to make STH even better.

With that said, we need help.

Hiring in Scottsdale

It is time to scale the team a bit. This quarter, we Welcomed the Newest STH Team Member, but he is taking time. My goal is to find someone who can become a #2 over time for STH. We also have some roles that we will be hiring for where the primary workplace will be in the STH studio.

Our team has been fantastic over the last five weeks. Despite everyone telling me that babies take a lot of time, experiencing it is different. We need help, or more specifically, I need help.

If you or someone you know wants to get into hardware reviews, drop me a line. Even if it is in a segment we do not currently focus on, feel free to pitch me on the idea.

The STH YouTube Update

This quarter, the growth of the STH YouTube channel again slowed a bit, but we reached 600K subscribers. Instead of doing a piece for every 100,000 (e.g. the 500,000 one), I think we will next mark 750,000.

My best guess is that we will crest 1,000,000 sometime in Q4 2025 if the growth rate does not slow too much. At the same time, that means we are thinking about evolving the channel for a 1M+ subscriber audience. We have a vision for that, but there is still a lot of ground to cover.

The Axautik Group Update

Last quarter, we rolled out our new analyst arm, the Axautik Group. This is the “Analyst” part of our Analyst-Influencer-Other model. Our first PCIe Gen6 and CXL retimer research short was more popular than I expected. At the same time, we have been working on getting some cool new research reports up at a lower price point, including a streamlined and updated piece on PCIe and CXL retimers. We will also have a few fascinating research reports in the networking and AI server space.

Aries Smart Retimer Add In Card For PCIe 5.0 Equinox
Aries Smart Retimer Add In Card For PCIe 5.0

The Axautik Group will produce content designed more for the financial communities. Some folks will be upset that AG content will be priced more for those communities. This is not a replacement for STH in any way. Instead, it will be designed to take much of what we learn by doing STH and turn it into formats we can monetize.

There is a question of whether this will be more of a subscription model or selling one-off reports.

STH Labs: The Shorts Channel Update

As a 2024 project, we have the new STH Labs shorts channel here.

While growth and posting frequency have been slower this quarter, we still managed to see the STH Labs shorts channel double. Realistically, it is still much smaller than the main channel, and we need to increase the volume of content here. My hope is that this crests 10,000 subscribers by early Q4.

Subscribe to the STH’s Newsletter and YouTube

Did you know STH has a weekly newsletter that comes out on Saturday with curated “Top 5” pieces from the week? We know you cannot visit every day, so we can deliver our picks for weekend reading directly to your inbox. Subscribing to the newsletter is easy. Here is the form.

Get the best of STH delivered weekly to your inbox. We are going to curate a selection of the best posts from STH each week and deliver them directly to you.

By opting-in you agree to have us send you our newsletter. We are using a third party service to manage subscriptions so you can unsubscribe at any time.

We are not selling your e-mail addresses and MailChimp is managing everything at this point so you can subscribe and unsubscribe from the list as you want.

We are not promoting the newsletter via overlays and pop-ups. Those are very effective, but they are bad for readers. I do not like them, so as long as I have a say, we are not going to have newsletter signup overlays. I run STH as something I would want to visit daily even if I did not work on it.

Finally, subscribe to our YouTube and check it out here. Since that is a big focus at this point.

Subscribing to STH’s newsletter helps you see my favorite pieces of each week and a preview of the next week. Subscribing to the STH YouTube channel also helps us demonstrate our reach beyond just the website. Sometimes, doing things like showing power consumption or fan noise is easier than writing about it. This has also been something important for STH over the past few months. We have to demonstrate reach, and simple things like subscriber counts help explain reach.

Final Words

Again, thank you to all of our readers and the STH team for making everything possible. We are on a mission to create the best content we can every day and present it in the best format we can every day. Over the next few months, there will be some growing pains, but we need to continue growing. Ultimately, the goal is for the best content we can produce every day to get better on average.

Hopefully, I pen the Q3 Letter at the end of September, and we can show more examples of how we want STH to evolve than I am articulating here.


  1. Great letter, thank you for the transparency that is so vital to legitimacy and longevity. It’s the only barrier left, honestly, and it’s the best kind. Been consuming STH in it’s various forms since it’s inception, and I’ll be here as long as you and/or your values are, Patrick. Thanks.

  2. Absolutely interested in helping out STH, i’m a sysadmin in Chandler and would love to get in contact

  3. What I’ve found by running my own business is that consumers place also zero premium on a company that has “values”. Consumers want quality, quantity, and low price (if there is a price). Don’t lose focus on what is making the site successful by attributing your success to your “values”. Having high ethical standards is nice, but what is making the site successful are more mundane things like frequent postings and detailed content. If you can’t keep that up, then it doesn’t matter what “values” you have because you won’t have customers to keep your business operating.

  4. I read every article about low-power/cost equipment, and at least skim every article about the higher-end stuff. I think you accurately outlined your strengths and the reasons you’ve built up an audience over the last few years: Fact-driven, independent content on a website that doesn’t suck.

    I assume you don’t feel substack/patreon are a good fit for the site since you didn’t mention them, but I think a large enough slice of your audience would be willing to pay on those platforms for the content you’re already making to make it worthwhile. I’m in for $5/m if you want to give it a shot.

  5. I just want to double down on what Leo said above. Patreon is the primary way I support the “media” sources I enjoy. I would gladly sign up for a $5/month membership to help keep STH the way we all love. Thanks Patrick for all you do and best wishes to your growing family.

  6. I’d argue that anyone that relies on AI output for their answers is a fool. Always read the source. AI gets shit COMPLETELY wrong all the time.

    Also, if AI kills off the source, then what are they going to train on going forward?

  7. “Also, if AI kills off the source, then what are they going to train on going forward?”

    I do agree, but this is the conundrum our culture is willing to go down….


Please enter your comment!
Please enter your name here

This site uses Akismet to reduce spam. Learn how your comment data is processed.