DRAFT - blog 2025 review from James #14256
Open
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
The latest updates on your projects. Learn more about Vercel for GitHub.
|
ivanagas
approved these changes
Jan 5, 2026
Contributor
ivanagas
left a comment
ivanagas
left a comment
There was a problem hiding this comment.
assorted tweaks, I can add some more internal links once you do another pass.
| featuredImageType: full | ||
| --- | ||
|
|
||
| It's December 31st, 2025, and I've decided to hang out in a near-empty WeWork to write about how the year went. I want to clear my head before next year starts, and thought it make for an interesting read for others. |
Contributor
There was a problem hiding this comment.
Suggested change
| It's December 31st, 2025, and I've decided to hang out in a near-empty WeWork to write about how the year went. I want to clear my head before next year starts, and thought it make for an interesting read for others. | |
| It's December 31st, 2025, and I've decided to hang out in a near-empty WeWork to write about how the year went. I want to clear my head before next year starts, and thought it'd make for an interesting read for others. |
|
|
||
| ## Money | ||
|
|
||
| Money is the outcome of doing things well for customers, in the long term. So let's take a closer look how 2025 went. |
Contributor
There was a problem hiding this comment.
Suggested change
| Money is the outcome of doing things well for customers, in the long term. So let's take a closer look how 2025 went. | |
| Money is the outcome of doing things well for customers in the long term. So let's take a closer look how 2025 went. |
|
|
||
| Money is the outcome of doing things well for customers, in the long term. So let's take a closer look how 2025 went. | ||
|
|
||
| We got super close to tripling our revenue and only burned about $5M in the process. We have reached around 285,000 companies having installed our software and 545,000 users. Somewhat amusingly, we said we wanted to hit $100M in 2026, back when we had only around $100K in annual revenue. Somehow that is still the case - we aim to hit $100M next year still! The internal codename for this project has become "the hundy", somehow. |
Contributor
There was a problem hiding this comment.
Suggested change
| We got super close to tripling our revenue and only burned about $5M in the process. We have reached around 285,000 companies having installed our software and 545,000 users. Somewhat amusingly, we said we wanted to hit $100M in 2026, back when we had only around $100K in annual revenue. Somehow that is still the case - we aim to hit $100M next year still! The internal codename for this project has become "the hundy", somehow. | |
| We got super close to tripling our revenue and only burned about $5M in the process. We reached around 285,000 companies having installed our software and 545,000 users. Somewhat amusingly, we said we wanted to hit $100M in 2026, back when we had only around $100K in annual revenue. Somehow that is still the case - we aim to hit $100M next year still! The internal codename for this project has become "the hundy", somehow. |
|
|
||
| This growth is outrageously fast, but the funny thing is that this doesn't sound amazing compared to other companies going from $0 to $100M in a year. This has genuinely been a challenge to handle - "is this FOMO or a good idea for _our_ users" is a criticism I've received this year. | ||
|
|
||
| Recently I went on a short trip to Dubai with some of our investors and I chatted about this phenomen with Arnav from PeakXV. He said he could see a world where people look at AI-inference heavy companies as a different business model to SAAS, where it's acceptable to have a gross margin 25% of what has been standard in SAAS, but thus you just need to grow 4 times faster. Net profit growth being the metric, not "ARR" growth. It'll be interesting to see how we handle this through 2026 as we ship a bunch of AI first products and features. |
Contributor
There was a problem hiding this comment.
This reads like "gross margin of 25%"
Suggested change
| Recently I went on a short trip to Dubai with some of our investors and I chatted about this phenomen with Arnav from PeakXV. He said he could see a world where people look at AI-inference heavy companies as a different business model to SAAS, where it's acceptable to have a gross margin 25% of what has been standard in SAAS, but thus you just need to grow 4 times faster. Net profit growth being the metric, not "ARR" growth. It'll be interesting to see how we handle this through 2026 as we ship a bunch of AI first products and features. | |
| Recently I went on a short trip to Dubai with some of our investors and I chatted about this phenomena with Arnav from PeakXV. He said he could see a world where people look at AI-inference heavy companies as a different business model to SaaS, where it's acceptable to have a gross margin a quarter of what has been standard in SaaS, thus you just need to grow 4 times faster. Net profit growth being the metric, not "ARR" growth. It'll be interesting to see how we handle this through 2026 as we ship a bunch of AI first products and features. |
|
|
||
| Recently I went on a short trip to Dubai with some of our investors and I chatted about this phenomen with Arnav from PeakXV. He said he could see a world where people look at AI-inference heavy companies as a different business model to SAAS, where it's acceptable to have a gross margin 25% of what has been standard in SAAS, but thus you just need to grow 4 times faster. Net profit growth being the metric, not "ARR" growth. It'll be interesting to see how we handle this through 2026 as we ship a bunch of AI first products and features. | ||
|
|
||
| Long ago we realized that our growth was largely from word of mouth. This comes out as a _lot_ of organic search volume, but mostly for the word "PostHog". We hired a friend and former colleague for a year or so and he went one level deeper - on finding out a user was recommended PostHog, he asked them _why_. They came back with four reasons: lots of products in one, the lowest prices, developer brand and a very technical and helpful customer service. We've been trying to do more and more of those things over time to grow. |
Contributor
There was a problem hiding this comment.
Suggested change
| Long ago we realized that our growth was largely from word of mouth. This comes out as a _lot_ of organic search volume, but mostly for the word "PostHog". We hired a friend and former colleague for a year or so and he went one level deeper - on finding out a user was recommended PostHog, he asked them _why_. They came back with four reasons: lots of products in one, the lowest prices, developer brand and a very technical and helpful customer service. We've been trying to do more and more of those things over time to grow. | |
| Long ago we realized that our growth was largely from word of mouth. This comes out as a _lot_ of organic search volume, but mostly for the word "PostHog". We hired a friend and former colleague for a year or so and he went one level deeper - on finding out a user was recommended PostHog, he asked them _why_. They came back with four reasons: lots of products in one, the lowest prices, developer brand, and very technical and helpful customer service. We've been trying to do more and more of those things over time to grow. |
Member
There was a problem hiding this comment.
I'm confused about who we hired for a year or so, why they left, and why that's relevant to this sentence.
|
|
||
| We raised a ton of money too - $145M of capital for the business through our Series D and E this year. Psychologically this has been a big shift. We wouldn't have known how to reliably spend money to grow in 2023, we started to get an idea through 2024 and this was the year we acted on it. It means that our constraint has shifted from confidence to capital to our ability to find really strong engineers. That's a pretty strange feeling and a little unnerving. It can feel frustrating at times - we have _so_ many things to build at once that we believe in. I am very glad we are able to hire across so many countries. I couldn't imagine trying to do this in just commuting distance of one or two offices. | ||
|
|
||
| One thing we are realizing about hiring - we get lots of applications for the roles that we market to. We haven't really done much outbound recruitment at all. We write all the time about product engineering - we have a [newsletter](https://newsletter.posthog.com/) on this topic with over 100K subscribers, and we build tools for product engineers. However, when it comes to finding different engineering expertise - such as deeper infrastructure stuff - applicants are understandably rarer. We want to figure out outbound recruitment in 2026 to make sure we get people with some of the more specialist skills we need as we keeping scaling our stack. |
Contributor
There was a problem hiding this comment.
I didn't quite understand what you were saying, but I think this makes more sense
Suggested change
| One thing we are realizing about hiring - we get lots of applications for the roles that we market to. We haven't really done much outbound recruitment at all. We write all the time about product engineering - we have a [newsletter](https://newsletter.posthog.com/) on this topic with over 100K subscribers, and we build tools for product engineers. However, when it comes to finding different engineering expertise - such as deeper infrastructure stuff - applicants are understandably rarer. We want to figure out outbound recruitment in 2026 to make sure we get people with some of the more specialist skills we need as we keeping scaling our stack. | |
| One thing we are realizing about hiring - we're going to need to do outbound recruitment for some roles. We traditionally haven't needed to do this as we market to the same audience we get applicants from - we build tools for product engineers and have a [newsletter](https://newsletter.posthog.com/) about this with over 100K subscribers. However, when it comes to finding different engineering expertise - such as deeper infrastructure stuff - applicants are understandably rarer. We want to figure out outbound recruitment in 2026 to make sure we get people with some of the more specialist skills we need as we keeping scaling our stack. |
|
|
||
| We got caught out by [Shai Hulud](https://posthog.com/blog/nov-24-shai-hulud-attack-post-mortem). We've always considered ourselves a data company. An attack that instead treated us as an initial vector for an ecosystem-wide worm wasn't something we'd anticipated. It was super frustrating this happened as we take security very seriously and in general today we have a very cautious attitude towards it. | ||
|
|
||
| ### PostHog AI is now awesome, but we launched PostHog AI too early, perhaps |
Contributor
There was a problem hiding this comment.
Not saying PostHog AI twice in one title
Suggested change
| ### PostHog AI is now awesome, but we launched PostHog AI too early, perhaps | |
| ### PostHog AI is now awesome, but we probably launched it too early |
|
|
||
| The mental model I used to hold of PostHog was similar to that I have of AWS - it is hundreds of products, you simply ignore 95% of them. They're all reliable and have most of the functionality of the competitors so you just pick them by default each time since you already have an account and it's often better to use them together. It's complex but you ignore most of it. | ||
|
|
||
| AI has changed this view - we've already demonstrated that agents can use our products together. My personal use of PostHog is now AI first. |
Contributor
There was a problem hiding this comment.
Trying to square the circle of why AI means simpler, there's something slightly confusing here.
Suggested change
| AI has changed this view - we've already demonstrated that agents can use our products together. My personal use of PostHog is now AI first. | |
| AI has changed this view - it means you will be able to use all our products together and be simpler. My personal use of PostHog is now AI first. |
|
|
||
| ## We think 2026 will be one of our most challenging years | ||
|
|
||
| 2026 is the year we go from being the underdog to being the same order of magnitude in terms of revenue. We have a far bigger customer base than many of them in terms of numbers. |
Contributor
There was a problem hiding this comment.
underdog to who? customer base compared to who? 😅
|
|
||
| 2026 is the year we go from being the underdog to being the same order of magnitude in terms of revenue. We have a far bigger customer base than many of them in terms of numbers. | ||
|
|
||
| We think the key is we have to get bigger at handling larger customers - everyone says you're a toy until you aren't! - _and_ we have to keep innovating and building furiously. We'd better get on with hiring then! |
Contributor
There was a problem hiding this comment.
Suggested change
| We think the key is we have to get bigger at handling larger customers - everyone says you're a toy until you aren't! - _and_ we have to keep innovating and building furiously. We'd better get on with hiring then! | |
| We think the key is we have to get bigger at handling larger customers - everyone says you're a toy until you aren't! - _and_ we have to keep innovating and building furiously. We'd better get on with hiring then! [Join us](/careers)! |
raquelmsmith
reviewed
Jan 6, 2026
|
|
||
| We kept up with our price cuts - reducing prices for warehouse, data pipeline and surveys customers. These hurt in the short run but lead to more growth long run for us. The cool thing is that these moves aren't driven directly by Tim and me any more - they're just part of [our principles](https://posthog.com/handbook/engineering/feature-pricing). These changes all come from small teams making the decision, making it happen then measuring the impact. | ||
|
|
||
| We shipped a ridiculous amount of stuff - far too much to list. New products (workflows, error tracking, logs, PostHog AI) and features, stability and performance improvements for our existing stack. And a fun new operating-system inspired website! |
Member
There was a problem hiding this comment.
Link the changelog?
Suggested change
| We shipped a ridiculous amount of stuff - far too much to list. New products (workflows, error tracking, logs, PostHog AI) and features, stability and performance improvements for our existing stack. And a fun new operating-system inspired website! | |
| We shipped a ridiculous amount of stuff - far too much to list here, but you can find most of it on the [changelog](/changelog). New products (workflows, error tracking, logs, PostHog AI) and features, stability and performance improvements for our existing stack. And a fun new operating-system inspired website! |
raquelmsmith
reviewed
Jan 6, 2026
|
|
||
| Then I took over a bunch of engineering teams. Then I dropped marketing. Now I just focus on AI products and the website. Tim is responsible for _everything_ else. We've swapped so much of our roles around. | ||
|
|
||
| This has worked remarkably well for us - despite having the "wrong" experience, we are both having much more fun and have way more energy. It turned out that what we're actually best at, wasn't what we had experience at doing initially. Don't default into doing what you've done before - pick what you want to focus more on. You'll figure it out and you'll have more fun regardless. Even today, everyone in a leadership role at PostHog was hired doing individual contributor work first, often for years. I think evolution means it's human nature to put people into boxes prematurely. |
Member
There was a problem hiding this comment.
Suggested change
| This has worked remarkably well for us - despite having the "wrong" experience, we are both having much more fun and have way more energy. It turned out that what we're actually best at, wasn't what we had experience at doing initially. Don't default into doing what you've done before - pick what you want to focus more on. You'll figure it out and you'll have more fun regardless. Even today, everyone in a leadership role at PostHog was hired doing individual contributor work first, often for years. I think evolution means it's human nature to put people into boxes prematurely. | |
| This has worked remarkably well for us - despite having the "wrong" experience, we are both having much more fun and have way more energy. It turned out that what we're actually best at wasn't what we had experience at doing initially. Don't default into doing what you've done before - pick what you want to focus more on. You'll figure it out and you'll have more fun regardless. Even today, everyone in a leadership role at PostHog was hired doing individual contributor work first, often for years. I think it's human nature to put people into boxes prematurely, but it's worth resisting. |
Member
|
Love this summary, James! |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Changes
Added reflections on how 2025 went. Hihg level feedback welcome.
Thought it might be nice to email to our user base as an investor update at some stage - maybe that needs to be thought through better? ideas around this also welcome.
Checklist
vercel.jsonArticle checklist