I am not sure why this old news is surfacing here today but I can give my 2 cents, since I sold speedchecker.com last year and were directly competing with Ookla.
The main business is selling the data. You use Speedtest.net to troubleshoot your connection but metrics captured with the test alongside location data give telcos invaluable insights on where they should improve their networks. Telcos pay 6 figures annually for this data and we have a few hundreds of of those big MNOs globally. This market is pretty big. Accenture is in trouble with their main consulting business due to AI so acquiring data business is one of the smart strategies they can implement to stay relevant.
To all commenters who think they can code it over the weekend, yes you are right. I coded my first speed checker over the weekend in 2008 but it took me 18 years to grow the user base , figure out entreprise sales strategy and exit. Its not easy as it seems.
As someone who worked at a CDN for years, I imagine the code is the easiest technical part. Managing the infrastructure, network connectivity, load balancing, and capacity planning would be the harder parts, outside of the sales and marketing bits.
If you don’t get all of those parts right, you are going to end up measuring your own bandwidth rather than the client’s.
The website, and backend code for the test. 10% of the software work. Which is what everyone seems to think.
The code to managing the infrastructure, network connectivity, load balancing, and capacity planning is the 90% of the software part. But even then it is only 10% of the technical thing.
Getting all the ISP onboard to have your server in their network / exchange and to deal with you, takes more time and effort then all the software part. But even then it is only 10% of the project.
The remaining 90%? Non technical part for Sales and Marketing and getting user traction.
To put that into perspective, the website can be done in a weekend was only 0.01% of the work.
> but it took me 18 years to grow the user base , figure out entreprise sales strategy and exit.
The audience here has never wanted to admit that the codebase doesn't really matter. Now that codebases can be created in a weekend, people are opening their eyes to this sentiment - the hard part is the sales, the code is easy.
Great code is still not easy. Choosing the right stack/libraries/billing and getting everything to work together (for cheap) is still something barely 10% of devs can actually realistically do.
Sales is hard, yeah, but look at everyone claiming to be building something amazing and it ends up 9 months behind schedule or just being an buggy, untested version of something that already exists in the market.
Great code has never been the requirement of success.
Throughout my career no software that hits $100m annual revenue was born from great code. That’s 2 fortune 100 hi-tech companies with other medium sized companies with revenue close to $1B.
There was one company that had better codebase than the others, unfortunately that company struggles to hit $2M MRR…
Looking back, it was painful to admit that code quality was not how the company succeed: it was overall strategy and luck.
Most of them can't. It's just a few products like speed testers and social media that are at least surface level easy. You can't vibe code up a triple A quality game in a weekend.
To be fair, triple A gaming lately has been something of a letdown. That does not invalidate your point exactly, but it does mean that the tastes of the gaming audience might be changing ( we kinda were in similar spots before ).
I’m sure the viewpoint from being in mergers and acquisitions is quite different (and to me, often comes across as quite callow). I’ve been a software developer for 35 years (closer to 45 if you include my pre-professional life, aka adolescence) and have deliberately stayed “on the tools” in my career with working in codebases and product development as I’ve found that is where I am happiest and can make the best contribution, rather than move up the managerial ladder to my level of incompetence, to quote Peter.
To create a successful product in IT, or any industry really, it takes a lot of different skills, facets and (often competing) priorities. And those priorities do change over time. I’m sure by the time a product or service crosses your desk, the codebase quality is not as big of a priority. Earlier in the life cycle a shit codebase makes for a shit product that is a lot harder to grow and maintain — so much so that most of them have probably folded before they reached the stage of looking to be merged or acquired. I’ve dabbled in sound mixing for live performance and when training others I’ve mentioned the fact that it very hard to make a bad singer or musician sound good, but very easy to make a good singer or musician sound bad. Same goes for trying to make what would otherwise have been a good product or service with a bad codebase. That’s really hard and creates a hell of a lot more work for every part of the business.
I’ve had sales people tell me to my face that they are the most important part of the business and the actual product or services is not that important. And in my more callow stages of life experience I’m pretty sure I’ve reciprocated with words like useless and parasitic, and that I could replace them with a small bash script. But in reality what we all do is important to the complex endeavour of developing and maintaining a successful product or service. The existential threat of AI is moving up the ladder of incompetence and changing the face of what we do. It may even jump a few rungs in the process. But it’s not there just yet. Keep making good sales, keep making good mergers, good products, good acquisitions, good services, and good codebases.
— No tokens were harmed in the production of this comment. —
Your argument is sound. It certainly takes a good deal of skill to create good code. And yes, good code makes it easier to create a better product.
And yes it's easier to build a better company on a better product.
But history is littered with "worse products" that won in the marketplace.
It turns out that all the attributes you name are helpful but not necessary. Good marketing trumps good product. We see this over and over again.
The best combination is good marketing and good product. If I can only get 1 of those then I'll take hood marketing. Equally if you have a good product but bad marketing you don't get many (if any) users. The "ask" section on this site is littered with that.
So, assuming we can all make "good enough" code, the code doesn't matter. It's all good enough. The distinguishing feature is the marketing, because that leads to market share, and that's all any company is really selling (once it sells for a lot).
I'm upvoting you because your comment is well made, and certainly common, even if it is incorrect:)
Having been involved in multiple different acquisitions, on both sides of the table, I can anecdote that the code quality had no impact on any part of the acquisitions. The players are not buying or selling the code.
Your entire argument hinges on "good enough". Problem is: you can never know if something is "good enough", except in hindsight for those products that succeeded.
I'm upvoting you because your comment is well made, and certainly common, even if it is nothing more than a tautology :)
I've been in an acquisition where code quality was important. But it was probably an edge case since the buyer just wanted to turn the company into a feature, and ease of integration into the buyer was important.
I think this perspective benefits from experience, the ability to step outside one’s self, see that the world is complicated, then focus on the thing you enjoy.
As much as I agree with you now, I also accept that younger me wouldn’t have!
> The audience here has never wanted to admit that the codebase doesn't really matter.
Are we talking about speed testing websites or the code that controls space vehicles? Perhaps extreme generalities do not provide useful insights.
> Now that codebases can be created in a weekend
Now that corporations are whitewashing copyright off of code so you can steal it without conscience.
> people are opening their eyes to this sentiment
Code is the product. Engineering is the discipline. That you can achieve high sales without good engineering is not a new idea. That it only provides short term benefits and leaves you irrelevant in the long term is the actual sentiment.
> the code is easy.
Coding has been easy since Perl was released. Knowing _what_ to code is the problem.
The partner network of Speedtest is also impressive. I don't know how many speed tests they need to handle in parallel, but usually it's always enough to do speed tests up to 5-10 Gbit/s. With more and more fiber connections also latency becomes very relevant. Otherwise the tests would be meaningless. Speedtest manages to measure less than 1ms latency on my fiber connection.
Once you have a good amount of users testing, its not that difficult to get free servers from the ISPs. The secret is that on-net servers show testers better performance than off-net so every ISP wants to contribute the speed test server. If they dont do it they are shooting themselves in the foot by routing their traffic to competitor networks and getting test results behind their peers.
Whats even worst then your competitors can claim awards for the Fastest ISP and your marketing people are furious!
That was a persistent conversation with ISPs when I was building a white label WiFi product (competitor to Eero).
Some ISPs wanted us to pin to their servers in our app to have the best possible results (we refused) while others wanted us to use their servers because they offered 10G service and none of the other servers had that much throughout. So their true 10G line would be limited by the server, not the line.
Sure, but it's still a huge effort to set up all those partnerships and keep them alive. ISPs are often traditional and slow moving companies, it probably takes a lot of work to get those servers in place at the right locations.
When you first built it, were you aware there was market for the data? Or was this something you discovered afterwards? It makes sense, but I wouldn’t have guessed it.
It’s less than ninety days old and it isn’t (2025), so I wouldn’t consider this as ‘old news’ yet. TIL, for example! But if you think it’s a dupe/repost and should be squashed, email the mods a link to both this and the prior post so they can evaluate.
I'm not sure about broadband data, as it can't be that useful. However on the mobile side, it's fairly valuable as a mobile app can collect A-GPS location and sensor telemetry that are unknown to the MNO otherwise.
Thank you for your service ( the product was/is -- haven't used in a while -- useful ).
I think this is the part that people do not appreciate. Sometimes it genuinely it is not the difficulty of the task from a pure programming perspective, but 1) getting the users and 2) getting people to pay for the service and 3) getting the right people to sign off on that.
It is very similar in banking. The products themselves are not super hard ( though the challenges are real ), buy just getting to talk to the right people is a hassle.
Are you able to share the details on how it was valued? Was it N times revenue or anything like that? I have tried to value a property several parties are interested in and found it quite mysterious.
thanks fornthe insights. inthought it was a wopping number but this makes totally sense. never realised this was gather valuable data for network operators. cool insights!
Any recommendations for similar tools to check network metrics other than speed ? Used to be a few free ones but would be nice to have an easy one to use
Gotta collect that data on how the services perform. lol
Snark aside, at least they have tight data controls and don’t sell that information.
But I don’t really have a lot of faith in the whole story and/or some of the management inside. Yeah there usually is always two sides to the coin. But tbh It’s too secretive. There has to be something there outside of simply “protecting the surprise.”
> I coded my first speed checker over the weekend in 2008 but it took me 18 years to grow the user base , figure out entreprise sales strategy and exit. Its not easy as it seems.
The biggest surprise (imo) when you start a business is how little of running a business is actually directly about the product. Having a product is essential, sure, and having a good product is nice, but that’s just the tip of the iceberg.
That test was totally inaccurate for me. It got the download right but upload was only 1/12 of my rated speed and 1/12 of what all the other tests (and my actual experience) tell me.
Cloudflare's undercounts my download by 25% and my upload by 75% (versus both speedtest.net and my observed sustained data rates). Also reports double the latency.
"By integrating Ookla’s data products, including Speedtest®, Downdetector®, Ekahau®, and RootMetrics®, Accenture will help Communications Service Providers (CSPs), hyperscalers, and enterprises optimize the mission-critical Wi-Fi and 5G networks that power their digital core. [...] Ookla’s data platform is anchored by more than 250 million consumer-initiated tests per month, complemented by controlled drive, walk, and embedded testing options"[1]
Is there some legal reason to scatter announcements with that many ® symbols, or do they just do it for style reasons / because they think it makes the announcement look more impressive?
i'm guessing that part of accenture's consulting business is helping people navigate the trademark registration process. so they've got to hype up the ®.
that's nuts, unless I'm missing something, it doesn't seem like those products are that mind blowingly complex... wow. Makes we want to try building my own for the hell of it.
Downdetector in fact just seems to be a website catalog with essentially a guestbook and hit counter...
Of course they are not complex. They do have a network effect though. If you go to your local ISP and say “hey, my 500mbps plan is only doing 100mbps on Speedtest.net”, they’ll “fix it” (usually by working with Ookla to put an edge endpoint on their network)
If you tell the “hey frankyspeeddetect.com isn’t doing my 500mbps” they’ll tell you to it’s an issue with that random website. ISPs and services reach out to Ookla to onboard with them because they have a network effect/mindshare of whatever you wanna call it
When I used a major cable ISP, often my connection seemed slow, so I'd go to speedtest.com. The speedtest would be fine... and then I would magically have faster network performance again.
It happened enough times that I'm suspicious the ISP had some way to detect if you run a speedtest, and then prioritized traffic to that customer.
This was one of the reasons given, at the time, for why Netflix created fast.com. It's served by the same infra that does their streaming, and is thus difficult for isps to game. That is, it'd be hard for them to do some hack to make fast.com numbers without also benefiting Netflix streaming performance in the bargain.
Actually I thought Netflix had already acquired Ookla / speedtest.com, so I was surprised to see this headline. But it looks like this was just the Mandela effect.
That said, why didn't Netflix acquire the market leader in this space? Creating their own seems way less useful, since network effects are the whole point.
Based on Accenture acquiring them, I’d guess the actual business wasn’t really interesting to Netflix. And that leaves the infrastructure, where the value they get is it being Netflix infrastructure. I can see why they spent the money on a really good brandable domain instead.
Because Netflix doesn't care what your connection to speedtest.net is, they care what your connection to your closest Netflix server box is. A while back, Comcast/your last-mile ISP was throttling traffic to Netflix to get Netflix to pay them. So while Netflix's box had plenty of bandwidth to their ISP, your ISP wasn't using it, intentionally. Fast.com was their response to that, so you could blame your ISP and not Netflix for being slow.
That’s a really over simplification of the issue. Plenty of Netflix edge CDNs are (and always were) ISP hosted. It’s a win-win for both and a complete no-brainer. The ISP v. Netflix argument was always about contract and margin negotiations. Flat rate, usage percentages, minimums, maximums, special plans, cuts, etc. who has the upper hand in the negotiation so to speak. Funnily enough the repeal of net neutrality gave those smaller ISPs much better position in the negotiation with big tech, not necessarily Comcast. The internet discord focused on Comcast and Verizon because fuck those guys. Who is gonna argue in favor of Comcast or Verizon? But the real winners were thousands of smaller regional ISPs.
The internet discourse focused on the big ISPs which refused to deploy CDN nodes and then said they needed to double-charge for peering capacity. Most smaller ISPs deployed those Open Connect nodes either becsuse they weren’t as greedy or felt that their customers had alternatives.
> It happened enough times that I'm suspicious the ISP had some way to detect if you run a speedtest, and then prioritized traffic to that customer.
ISPs definitely know when you run a speedtest.net test. 90% of the time, the data for that comes from boxes/services they host themselves. It’s not exactly hidden either. It’s a typical program any ISP can sign up for and you can easily see the destination the test is running against. I won’t be surprised if some have some logic to prioritize particular subscribers plan once they have detected a test from them. They probably view it as a “customer support calls reduction” feature.
>When I used a major cable ISP, often my connection seemed slow, so I'd go to speedtest.com. The speedtest would be fine... and then I would magically have faster network performance again.
Yeah, I suspect you could script it to do it daily. They definitely seem to deprioritize traffic from people that don't complain.
All these numbers are fake. They are all special cased in most ISPs with the cooperation of cloudflare, Netflix, OOkla, Akamai, Google, etc. The centralization of the internet around AWS, Azure, Google, Netflix, Cloudflare, etc has been a godsend to ISPs and the internet infrastructure in general. Maintaining good network conditions to 4 or 5 dozen networks and working with them closely is so so much easier than maintaining full peer-to-peer network conditions. Go ahead and try to test internet speed to your home network over a wireguard VPN and compare it to the performance of the same VPN when connecting to any of the major services. Try to setup a tunnel between your house and your friends house in the same city and test the speed and compare it to fast.com or cloudflare.com or speedtest.net
I had not heard of http://speed.cloudflare.com either. I just tried it and I did not get accurate numbers. wifiman.com, from Ubiquiti/Unifi team does provide more accurate numbers. fast.com numbers are pretty accurate as well.
I'm a huge fan of https://speed.cloudflare.com/ and you'll have to come with better evidence. Also fast.com doesn't even give upload speed and latency.
That's why speedtest.net is a great purchase for Accenture. Of course Cloudflare's speed test is accurate: it's a test of how fast your connection is to their network. No more, no less. That their network doesn't have the same PoPs means it'll have different numbers than Ookla's test, your ISPs advertised numbers, Netflix's test, and any other speed test. But for people that don't see the Internet as a pile of different interconnected networks, the conclusion that a particular test is inaccurate is a win for Accenture.
the best part is Downdetector is inaccurate as hell - if AWS is genuinely down, folks get curious and search other providers, causing Downdetector to mark them as down too
When you search for a service you get the current status and you get the option to report a problem.
The minimal you expect from such a service is to keep track of how many % of users are searching also reports an error. There might of course still be errors but that alone surely can't be it. But please correct me if I'm wrong.
Ookla has huge amounts of data, speedtest’s software is integrated into networks and used by hundreds of millions of users, they have the most comprehensive information about internet connections. You can recreate the software but you can’t recreate the data without decades of integration into what seems like every network.
https://www.ookla.com/ You can see an overview of the data they collect and sell on the corporate website
There is also the Italian one, AFAIK you can use this as an official tool to check whether your speed is at least the minimum speed that should be provided by contract.
I don't know whether it pings to italy even outside italy/eu
We used fast.com to speed test our new office internet connection and the next day got an irate email from corporate (who had argued we didn't need the new connection) about "watching Netflix all day". I imagine some C-level thought they had a real gotcha! moment until I showed them the site.
I read a while ago that certain ISPs will optimize the traffic to Netflix's servers, and so when you run fast.com (which is my default, by the way), you get your Internet speed for watching Netflix, but not necessarily for other things.
That was very relevant in some scenarios. When Spectrum was fighting with Netflix, they would force Netflix traffic to a peer circuit that was under provisioned as a shakedown tactic.
Fast.com would detect that, and you could bypass that nonsense by changing your DNS.
Out of curiosity, I just compared my home wifi between fast.com and cloudflare's speed tedt and got similar results, completely and definitely disproving (n=1) my claim above.
While neat that a government operates this, I’m not sure it’s a viable alternative for most users given that the servers are AFAIK all in Norway. For example, the latency from my network was 150-200ms (compared to 6ms for the Speedtest.net server) and the speed test results appear less consistent than they may be in/near Norway.
>A fartlek (Swedish for "speed play") is an unstructured, continuous running workout that mixes fast bursts of intense running with slower, recovery jogging.
Worked there for half a decade and helped a little on this deal but exited right before.
Like another commenter pointed out, the deal is a data acquisition. Ookla is multimillion dollar business thanks to its awards and data programs with almost every telco a customer. Accenture was already a competitor thanks to their Umlaut acquisition
For most consumers, Ookla = Speedtest but there’s a lot more beneath the surface. Ookla owns a drive-testing firm, Downdetector (consumer based outage reporting) and a thriving SDK & server network. Most of the data comes via background tests and embedded SDK tests.
I don’t think I would trust Downdetector in the hands of a company that its main business is consulting some of the same business to assess.
Imagine a large Accenture business being down. Would they provide that evidence even when that could harm their own SLA commitments with their clients?
I would never have had SpeedTest on my board of unicorns…that’s an unbelievable sale price. To all the agents who negotiated that deal, my hat is off to you.
They are embedded in several points in many, many networks all over the world. They get real-world metrics, sometimes live as events are happening. And they don't own most of this infra, it's hosted voluntarily inside service provider and corporate networks.
That seems like a lot for name recognition. I bet you could rebuild their technology for like $20m at the most, and buy 100% market share for like $100m easy. Unless they have some other assets other than the obvious?
Isn't Speedtest's huge dataset of Internet speeds mapped to time, location and IP address, as well as data on VPN usage (a user checks the speed of his/her direct connection then turns VPN on and checks over that too, all within the same session) such an asset?
I doubt they didn't collect all of that.
P.S. Now marry that huge dataset with services that Accenture provides, among others:
"In February 2025, Vice News spoke to a former Accenture employee under the condition of anonymity. His project on the WhatsApp team for Meta required him to sift through images and decide whether or not they depicted child sexual abuse, which he coped with "through a lot of substance abuse". The former employee claimed to have witnessed multiple missed opportunities to protect children, and alleged that one colleague had previously been arrested for possessing child abuse materials. In a statement, Accenture said they are "committed to helping companies keep their platforms safe through services such as content, advertising, and compliance reviews."¹
Fast.com has existed for 15 years yet isn't nearly as popular. It's easy to build a new speed test, but much harder to get people to use it.
Downdetector wins because of SEO. Most people don't get there directly, they google for "is $x down" and then get sent to downdecetor. Which from my understanding works by simply showing you how many people came to their site with those search terms. They don't actually check the sites.
Fast is a Netflix product so the fact that you've even heard of it is in direct relation to the weight of the brand that launched it.
speedtest.net has been the first search result on Google for "speed test" for decades. Partly the boost of domain SEO and partly the boost of it being an effective exit node for searches for that term for that long.
(Nobody searches "ookla" and nobody is going to search your tier-3 .com)
There is something I never liked about Crapcenture - it's corporate culture is so weird and almost cult-like. There is no doubt they are successful but I question whether that model should even exist in the first place.
I am not sure why this old news is surfacing here today but I can give my 2 cents, since I sold speedchecker.com last year and were directly competing with Ookla.
The main business is selling the data. You use Speedtest.net to troubleshoot your connection but metrics captured with the test alongside location data give telcos invaluable insights on where they should improve their networks. Telcos pay 6 figures annually for this data and we have a few hundreds of of those big MNOs globally. This market is pretty big. Accenture is in trouble with their main consulting business due to AI so acquiring data business is one of the smart strategies they can implement to stay relevant.
To all commenters who think they can code it over the weekend, yes you are right. I coded my first speed checker over the weekend in 2008 but it took me 18 years to grow the user base , figure out entreprise sales strategy and exit. Its not easy as it seems.
As someone who worked at a CDN for years, I imagine the code is the easiest technical part. Managing the infrastructure, network connectivity, load balancing, and capacity planning would be the harder parts, outside of the sales and marketing bits.
If you don’t get all of those parts right, you are going to end up measuring your own bandwidth rather than the client’s.
This.
The website, and backend code for the test. 10% of the software work. Which is what everyone seems to think.
The code to managing the infrastructure, network connectivity, load balancing, and capacity planning is the 90% of the software part. But even then it is only 10% of the technical thing.
Getting all the ISP onboard to have your server in their network / exchange and to deal with you, takes more time and effort then all the software part. But even then it is only 10% of the project.
The remaining 90%? Non technical part for Sales and Marketing and getting user traction.
To put that into perspective, the website can be done in a weekend was only 0.01% of the work.
> but it took me 18 years to grow the user base , figure out entreprise sales strategy and exit.
The audience here has never wanted to admit that the codebase doesn't really matter. Now that codebases can be created in a weekend, people are opening their eyes to this sentiment - the hard part is the sales, the code is easy.
Great code is still not easy. Choosing the right stack/libraries/billing and getting everything to work together (for cheap) is still something barely 10% of devs can actually realistically do.
Sales is hard, yeah, but look at everyone claiming to be building something amazing and it ends up 9 months behind schedule or just being an buggy, untested version of something that already exists in the market.
Great code has never been the requirement of success.
Throughout my career no software that hits $100m annual revenue was born from great code. That’s 2 fortune 100 hi-tech companies with other medium sized companies with revenue close to $1B.
There was one company that had better codebase than the others, unfortunately that company struggles to hit $2M MRR…
Looking back, it was painful to admit that code quality was not how the company succeed: it was overall strategy and luck.
Most of them can't. It's just a few products like speed testers and social media that are at least surface level easy. You can't vibe code up a triple A quality game in a weekend.
To be fair, triple A gaming lately has been something of a letdown. That does not invalidate your point exactly, but it does mean that the tastes of the gaming audience might be changing ( we kinda were in similar spots before ).
> the codebase doesn't really matter
Sigh.
I’m sure the viewpoint from being in mergers and acquisitions is quite different (and to me, often comes across as quite callow). I’ve been a software developer for 35 years (closer to 45 if you include my pre-professional life, aka adolescence) and have deliberately stayed “on the tools” in my career with working in codebases and product development as I’ve found that is where I am happiest and can make the best contribution, rather than move up the managerial ladder to my level of incompetence, to quote Peter.
To create a successful product in IT, or any industry really, it takes a lot of different skills, facets and (often competing) priorities. And those priorities do change over time. I’m sure by the time a product or service crosses your desk, the codebase quality is not as big of a priority. Earlier in the life cycle a shit codebase makes for a shit product that is a lot harder to grow and maintain — so much so that most of them have probably folded before they reached the stage of looking to be merged or acquired. I’ve dabbled in sound mixing for live performance and when training others I’ve mentioned the fact that it very hard to make a bad singer or musician sound good, but very easy to make a good singer or musician sound bad. Same goes for trying to make what would otherwise have been a good product or service with a bad codebase. That’s really hard and creates a hell of a lot more work for every part of the business.
I’ve had sales people tell me to my face that they are the most important part of the business and the actual product or services is not that important. And in my more callow stages of life experience I’m pretty sure I’ve reciprocated with words like useless and parasitic, and that I could replace them with a small bash script. But in reality what we all do is important to the complex endeavour of developing and maintaining a successful product or service. The existential threat of AI is moving up the ladder of incompetence and changing the face of what we do. It may even jump a few rungs in the process. But it’s not there just yet. Keep making good sales, keep making good mergers, good products, good acquisitions, good services, and good codebases.
— No tokens were harmed in the production of this comment. —
Your argument is sound. It certainly takes a good deal of skill to create good code. And yes, good code makes it easier to create a better product.
And yes it's easier to build a better company on a better product.
But history is littered with "worse products" that won in the marketplace.
It turns out that all the attributes you name are helpful but not necessary. Good marketing trumps good product. We see this over and over again.
The best combination is good marketing and good product. If I can only get 1 of those then I'll take hood marketing. Equally if you have a good product but bad marketing you don't get many (if any) users. The "ask" section on this site is littered with that.
So, assuming we can all make "good enough" code, the code doesn't matter. It's all good enough. The distinguishing feature is the marketing, because that leads to market share, and that's all any company is really selling (once it sells for a lot).
I'm upvoting you because your comment is well made, and certainly common, even if it is incorrect:)
Having been involved in multiple different acquisitions, on both sides of the table, I can anecdote that the code quality had no impact on any part of the acquisitions. The players are not buying or selling the code.
> So, assuming we can all make "good enough" code
Your entire argument hinges on "good enough". Problem is: you can never know if something is "good enough", except in hindsight for those products that succeeded.
I'm upvoting you because your comment is well made, and certainly common, even if it is nothing more than a tautology :)
But doesn't your counter point assume all products that failed were not good enough?
I've been in an acquisition where code quality was important. But it was probably an edge case since the buyer just wanted to turn the company into a feature, and ease of integration into the buyer was important.
Marketing without a product is called a scam. Marketing with a "good enough" product to sell it is the same.
Sigh.
Bless your heart.
I think this perspective benefits from experience, the ability to step outside one’s self, see that the world is complicated, then focus on the thing you enjoy.
As much as I agree with you now, I also accept that younger me wouldn’t have!
Very well said.
I'm not sure anyone was operating under the idea that a speed test website's code was the hard part.
> The audience here has never wanted to admit that the codebase doesn't really matter.
Are we talking about speed testing websites or the code that controls space vehicles? Perhaps extreme generalities do not provide useful insights.
> Now that codebases can be created in a weekend
Now that corporations are whitewashing copyright off of code so you can steal it without conscience.
> people are opening their eyes to this sentiment
Code is the product. Engineering is the discipline. That you can achieve high sales without good engineering is not a new idea. That it only provides short term benefits and leaves you irrelevant in the long term is the actual sentiment.
> the code is easy.
Coding has been easy since Perl was released. Knowing _what_ to code is the problem.
> Engineering is the discipline.
And even that is very rare in the field of software development.
The partner network of Speedtest is also impressive. I don't know how many speed tests they need to handle in parallel, but usually it's always enough to do speed tests up to 5-10 Gbit/s. With more and more fiber connections also latency becomes very relevant. Otherwise the tests would be meaningless. Speedtest manages to measure less than 1ms latency on my fiber connection.
Once you have a good amount of users testing, its not that difficult to get free servers from the ISPs. The secret is that on-net servers show testers better performance than off-net so every ISP wants to contribute the speed test server. If they dont do it they are shooting themselves in the foot by routing their traffic to competitor networks and getting test results behind their peers.
Whats even worst then your competitors can claim awards for the Fastest ISP and your marketing people are furious!
That was a persistent conversation with ISPs when I was building a white label WiFi product (competitor to Eero).
Some ISPs wanted us to pin to their servers in our app to have the best possible results (we refused) while others wanted us to use their servers because they offered 10G service and none of the other servers had that much throughout. So their true 10G line would be limited by the server, not the line.
Sure, but it's still a huge effort to set up all those partnerships and keep them alive. ISPs are often traditional and slow moving companies, it probably takes a lot of work to get those servers in place at the right locations.
And some cheeky traffic shaping by your competitors to make your results look worse than they should be.
When you first built it, were you aware there was market for the data? Or was this something you discovered afterwards? It makes sense, but I wouldn’t have guessed it.
It’s less than ninety days old and it isn’t (2025), so I wouldn’t consider this as ‘old news’ yet. TIL, for example! But if you think it’s a dupe/repost and should be squashed, email the mods a link to both this and the prior post so they can evaluate.
Telcos wanting to improve their networks is news to me. I always thought providing the bare minimum is basically their business model.
Providing the bare minimum when you're not is improving. ;)
How come ISPs aren't providing that data internally from observing their own traffic ?
I'm sure they do, but data about the speeds of other ISPs is also valuable.
I'm not sure about broadband data, as it can't be that useful. However on the mobile side, it's fairly valuable as a mobile app can collect A-GPS location and sensor telemetry that are unknown to the MNO otherwise.
Thank you for your service ( the product was/is -- haven't used in a while -- useful ).
I think this is the part that people do not appreciate. Sometimes it genuinely it is not the difficulty of the task from a pure programming perspective, but 1) getting the users and 2) getting people to pay for the service and 3) getting the right people to sign off on that.
It is very similar in banking. The products themselves are not super hard ( though the challenges are real ), buy just getting to talk to the right people is a hassle.
ISP seems to give higher network priority to Ookla so I'm not sure how useful it is compared to actual experience.
Like all speed tests it should be considered the maximum you can achieve.
@forcer would love to know how you "grow the user base , figure out entreprise sales strategy and exit"
Apparently it takes roughly the same amount of time as raising a child
Are you able to share the details on how it was valued? Was it N times revenue or anything like that? I have tried to value a property several parties are interested in and found it quite mysterious.
thanks fornthe insights. inthought it was a wopping number but this makes totally sense. never realised this was gather valuable data for network operators. cool insights!
Any recommendations for similar tools to check network metrics other than speed ? Used to be a few free ones but would be nice to have an easy one to use
Most will show you speed / latency / packetloss with the toggleable option for multi-single threaded.
What would you like to see?
> Accenture is in trouble with their main consulting business due to AI
Is this something being seen across all outsourcers like Accenture, Wipro, Infosys etc?
> To all commenters who think they can code it over the weekend, yes you are right... but it took me 18 years to grow the user base
I should get this printed and framed
MNOs => Mobile Network Operators
Did you ever fear that any big player like Google, Apple, Valve (though fewer mobiles), Meta, OpenAI (nowadays), etc decide to get involved?
Apple has their own internal speed app due to privacy concerns.
And on macOS Apple has `networkQuality` on the command line which runs a speed test against presumably the iCloud servers
Gotta collect that data on how the services perform. lol
Snark aside, at least they have tight data controls and don’t sell that information.
But I don’t really have a lot of faith in the whole story and/or some of the management inside. Yeah there usually is always two sides to the coin. But tbh It’s too secretive. There has to be something there outside of simply “protecting the surprise.”
https://theapplewiki.com/wiki/Speed
> I coded my first speed checker over the weekend in 2008 but it took me 18 years to grow the user base , figure out entreprise sales strategy and exit. Its not easy as it seems.
The biggest surprise (imo) when you start a business is how little of running a business is actually directly about the product. Having a product is essential, sure, and having a good product is nice, but that’s just the tip of the iceberg.
what was the point where you had enough data to make it worthwhile for telcos?
I would say if you can reach 1% of the population in a given country every month then you are starting to be interesting for the telcos.
What about city specific data? Is that equally/more/less valuable? At what scale do you think?
All it takes to defeat the business model is https://openspeedtest.com/
That test was totally inaccurate for me. It got the download right but upload was only 1/12 of my rated speed and 1/12 of what all the other tests (and my actual experience) tell me.
Same for me. The download speed was fine, but upload speed was 1.75x the actual one.
Same for me: download was okay, upload was completely wrong
My very limited experience in this space is that measuring uploads is actually quite a bit trickier and harder to nail down than download speeds
Or https://speed.cloudflare.com/
Cloudflare's undercounts my download by 25% and my upload by 75% (versus both speedtest.net and my observed sustained data rates). Also reports double the latency.
evidently putting “open” in the name of a website doesn’t suddenly make it not-bad.
Huh, interesting, I thought ISP would get enough metrics from their own devices
Just wanted to say congratulations!
Why did you sell it?
Because I got a good offer :)
Many such cases
Money without having to do more work > money that you have to keep working for
"By integrating Ookla’s data products, including Speedtest®, Downdetector®, Ekahau®, and RootMetrics®, Accenture will help Communications Service Providers (CSPs), hyperscalers, and enterprises optimize the mission-critical Wi-Fi and 5G networks that power their digital core. [...] Ookla’s data platform is anchored by more than 250 million consumer-initiated tests per month, complemented by controlled drive, walk, and embedded testing options"[1]
[1] https://newsroom.accenture.com/news/2026/accenture-to-acquir...
Is there some legal reason to scatter announcements with that many ® symbols, or do they just do it for style reasons / because they think it makes the announcement look more impressive?
Using the symbol allows them greater protection under the Lanham act, because it counts as “notice” that the mark is registered.
Without it, it limits your ability to recover damages from infringement.
When you are making the absurd case you’ve trademarked “speed test”, yes, you have to take pains to mark it.
i'm guessing that part of accenture's consulting business is helping people navigate the trademark registration process. so they've got to hype up the ®.
Legal. Gotta protect your trademarks.
To a degree, but I've never seen anything requiring you to show a mark literally every time it's mentioned.
that's nuts, unless I'm missing something, it doesn't seem like those products are that mind blowingly complex... wow. Makes we want to try building my own for the hell of it.
Downdetector in fact just seems to be a website catalog with essentially a guestbook and hit counter...
Of course they are not complex. They do have a network effect though. If you go to your local ISP and say “hey, my 500mbps plan is only doing 100mbps on Speedtest.net”, they’ll “fix it” (usually by working with Ookla to put an edge endpoint on their network)
If you tell the “hey frankyspeeddetect.com isn’t doing my 500mbps” they’ll tell you to it’s an issue with that random website. ISPs and services reach out to Ookla to onboard with them because they have a network effect/mindshare of whatever you wanna call it
When I used a major cable ISP, often my connection seemed slow, so I'd go to speedtest.com. The speedtest would be fine... and then I would magically have faster network performance again.
It happened enough times that I'm suspicious the ISP had some way to detect if you run a speedtest, and then prioritized traffic to that customer.
This was one of the reasons given, at the time, for why Netflix created fast.com. It's served by the same infra that does their streaming, and is thus difficult for isps to game. That is, it'd be hard for them to do some hack to make fast.com numbers without also benefiting Netflix streaming performance in the bargain.
Actually I thought Netflix had already acquired Ookla / speedtest.com, so I was surprised to see this headline. But it looks like this was just the Mandela effect.
That said, why didn't Netflix acquire the market leader in this space? Creating their own seems way less useful, since network effects are the whole point.
Based on Accenture acquiring them, I’d guess the actual business wasn’t really interesting to Netflix. And that leaves the infrastructure, where the value they get is it being Netflix infrastructure. I can see why they spent the money on a really good brandable domain instead.
Because Netflix doesn't care what your connection to speedtest.net is, they care what your connection to your closest Netflix server box is. A while back, Comcast/your last-mile ISP was throttling traffic to Netflix to get Netflix to pay them. So while Netflix's box had plenty of bandwidth to their ISP, your ISP wasn't using it, intentionally. Fast.com was their response to that, so you could blame your ISP and not Netflix for being slow.
That’s a really over simplification of the issue. Plenty of Netflix edge CDNs are (and always were) ISP hosted. It’s a win-win for both and a complete no-brainer. The ISP v. Netflix argument was always about contract and margin negotiations. Flat rate, usage percentages, minimums, maximums, special plans, cuts, etc. who has the upper hand in the negotiation so to speak. Funnily enough the repeal of net neutrality gave those smaller ISPs much better position in the negotiation with big tech, not necessarily Comcast. The internet discord focused on Comcast and Verizon because fuck those guys. Who is gonna argue in favor of Comcast or Verizon? But the real winners were thousands of smaller regional ISPs.
The internet discourse focused on the big ISPs which refused to deploy CDN nodes and then said they needed to double-charge for peering capacity. Most smaller ISPs deployed those Open Connect nodes either becsuse they weren’t as greedy or felt that their customers had alternatives.
I meant, acquire the speedtest.net domain and point it to servers inside Netflix' farm.
> It happened enough times that I'm suspicious the ISP had some way to detect if you run a speedtest, and then prioritized traffic to that customer.
ISPs definitely know when you run a speedtest.net test. 90% of the time, the data for that comes from boxes/services they host themselves. It’s not exactly hidden either. It’s a typical program any ISP can sign up for and you can easily see the destination the test is running against. I won’t be surprised if some have some logic to prioritize particular subscribers plan once they have detected a test from them. They probably view it as a “customer support calls reduction” feature.
>When I used a major cable ISP, often my connection seemed slow, so I'd go to speedtest.com. The speedtest would be fine... and then I would magically have faster network performance again.
Yeah, I suspect you could script it to do it daily. They definitely seem to deprioritize traffic from people that don't complain.
http://speed.cloudflare.com is a bit harder to argue with though.
All these numbers are fake. They are all special cased in most ISPs with the cooperation of cloudflare, Netflix, OOkla, Akamai, Google, etc. The centralization of the internet around AWS, Azure, Google, Netflix, Cloudflare, etc has been a godsend to ISPs and the internet infrastructure in general. Maintaining good network conditions to 4 or 5 dozen networks and working with them closely is so so much easier than maintaining full peer-to-peer network conditions. Go ahead and try to test internet speed to your home network over a wireguard VPN and compare it to the performance of the same VPN when connecting to any of the major services. Try to setup a tunnel between your house and your friends house in the same city and test the speed and compare it to fast.com or cloudflare.com or speedtest.net
I had not heard of http://speed.cloudflare.com either. I just tried it and I did not get accurate numbers. wifiman.com, from Ubiquiti/Unifi team does provide more accurate numbers. fast.com numbers are pretty accurate as well.
I'm a huge fan of https://speed.cloudflare.com/ and you'll have to come with better evidence. Also fast.com doesn't even give upload speed and latency.
Sure it does.
Just tap ‘More Info’ to show them
That's not the point of fast.com
Fast is a one click solution to finding out your download speed from Netflix.
Latency doesn't matter, nor does upload.
They serve so many sites that they are probably the best test there's now.
> I did not get accurate numbers.
That's why speedtest.net is a great purchase for Accenture. Of course Cloudflare's speed test is accurate: it's a test of how fast your connection is to their network. No more, no less. That their network doesn't have the same PoPs means it'll have different numbers than Ookla's test, your ISPs advertised numbers, Netflix's test, and any other speed test. But for people that don't see the Internet as a pile of different interconnected networks, the conclusion that a particular test is inaccurate is a win for Accenture.
The difference is that until now I had never heard of speed.cloudflare.com before. (I know about fast.com though.)
The valuation must be outside of the tech. Are there relationships or contracts Accenture is getting access to?
Clearly. They are buying all the servers that are already imbedded in ISP networks.
or overlapping board members who are essentially paying themselves
it must be
thefacebook.com was developed in a few days, too. The value was never in the code.
the best part is Downdetector is inaccurate as hell - if AWS is genuinely down, folks get curious and search other providers, causing Downdetector to mark them as down too
Just searching surely does not mark them as down?
That's exactly how it works. The website is just a counter of how many people landed on the page searching for "is $X down"
I don't know but I really doubt that.
When you search for a service you get the current status and you get the option to report a problem.
The minimal you expect from such a service is to keep track of how many % of users are searching also reports an error. There might of course still be errors but that alone surely can't be it. But please correct me if I'm wrong.
edit: and their own description only mentions actual reports https://downdetector.com/methodology/
It does.
dont miss it, its almost all about users and revenue not how complex or simple product is.
Ookla has huge amounts of data, speedtest’s software is integrated into networks and used by hundreds of millions of users, they have the most comprehensive information about internet connections. You can recreate the software but you can’t recreate the data without decades of integration into what seems like every network.
https://www.ookla.com/ You can see an overview of the data they collect and sell on the corporate website
That's a lot of money just for network topography, but may someone let them in and it has a whole map of an otherwise hidden or inaccessible network.
You’re not paying for the tech, you’re paying for the name and the users. Speedyest claims 40m unique users per month.
The fact that you think that the code is all that material here is exactly why you aren’t the person to be doing this. Go right ahead.
ZIR SV culture has built multiple generations of nerds that think that they can just effortlessly become billionaires. Ridiculous brain rot.
speedtest has a lot of volunteers hosting local servers, which you need to do a good last mile speed test.
that capilarity is not something you can achieve overnight.
The valuation is the name recognition and that that’s where people go to do those things
I really hope they don't change the interface of the speedchecker because it's so clean rn
may i suggest nettfart.no by the norwegian government as an alternative ? at least the name is fun
It's not the net fart that kills your connection, it is the server smell.
There is also the Italian one, AFAIK you can use this as an official tool to check whether your speed is at least the minimum speed that should be provided by contract.
I don't know whether it pings to italy even outside italy/eu
https://misurainternet.it/misura-speedtest/?speedtest=inizia
fast.com is my go-to in the rare case I need to check network speed these days
We used fast.com to speed test our new office internet connection and the next day got an irate email from corporate (who had argued we didn't need the new connection) about "watching Netflix all day". I imagine some C-level thought they had a real gotcha! moment until I showed them the site.
This is another advantage of fast.com.
I read a while ago that certain ISPs will optimize the traffic to Netflix's servers, and so when you run fast.com (which is my default, by the way), you get your Internet speed for watching Netflix, but not necessarily for other things.
The opposite can also be true. T-Mobile throttles Netflix and fast.com on my 5g mobile plan to be less than 5Mbps where Speedtest shows > 200Mbps.
That was very relevant in some scenarios. When Spectrum was fighting with Netflix, they would force Netflix traffic to a peer circuit that was under provisioned as a shakedown tactic.
Fast.com would detect that, and you could bypass that nonsense by changing your DNS.
Out of curiosity, I just compared my home wifi between fast.com and cloudflare's speed tedt and got similar results, completely and definitely disproving (n=1) my claim above.
For Swedish users: <https://www.bredbandskollen.se/>
Finnish https://bittimittari.fi/en/measurement
While neat that a government operates this, I’m not sure it’s a viable alternative for most users given that the servers are AFAIK all in Norway. For example, the latency from my network was 150-200ms (compared to 6ms for the Speedtest.net server) and the speed test results appear less consistent than they may be in/near Norway.
why is it named this? i'm guessing "fart" means something different in your language :-)
"speed" in Swedish and Norwegian. Probably Danish as well.
ah, makes sense, thank you.
>A fartlek (Swedish for "speed play") is an unstructured, continuous running workout that mixes fast bursts of intense running with slower, recovery jogging.
Wait until you find out what they call "closed".
I recommend https://speed.cloudflare.com/ personally.
I like the fact that it shows packet loss
Well, I originally found it on the Bufferfloat test page: https://www.bufferbloat.net/projects/bloat/wiki/Tests_for_Bu...
Worked there for half a decade and helped a little on this deal but exited right before.
Like another commenter pointed out, the deal is a data acquisition. Ookla is multimillion dollar business thanks to its awards and data programs with almost every telco a customer. Accenture was already a competitor thanks to their Umlaut acquisition
For most consumers, Ookla = Speedtest but there’s a lot more beneath the surface. Ookla owns a drive-testing firm, Downdetector (consumer based outage reporting) and a thriving SDK & server network. Most of the data comes via background tests and embedded SDK tests.
May it fail and take Accenture with them.
Accenture is more resilient than cockroaches.
I don’t think I would trust Downdetector in the hands of a company that its main business is consulting some of the same business to assess.
Imagine a large Accenture business being down. Would they provide that evidence even when that could harm their own SLA commitments with their clients?
I would trust Datadog more with https://updog.ai/
Big fan of https://speedtest.aa.net.uk
Terrible domain name.
Not THAT Ookla, darn it. http://www.ooklathemok.com/
I would never have had SpeedTest on my board of unicorns…that’s an unbelievable sale price. To all the agents who negotiated that deal, my hat is off to you.
They are embedded in several points in many, many networks all over the world. They get real-world metrics, sometimes live as events are happening. And they don't own most of this infra, it's hosted voluntarily inside service provider and corporate networks.
And now every SLA that Accenture is held to for uptime suddenly will never be breached…
use https://speedof.me
https://speed.cloudflare.com
Yes.
Why tho?
It's almost always user data these days, so probably that.
Maybe the owners were ready to cash out and move on with life?
That seems like a lot for name recognition. I bet you could rebuild their technology for like $20m at the most, and buy 100% market share for like $100m easy. Unless they have some other assets other than the obvious?
Isn't Speedtest's huge dataset of Internet speeds mapped to time, location and IP address, as well as data on VPN usage (a user checks the speed of his/her direct connection then turns VPN on and checks over that too, all within the same session) such an asset?
I doubt they didn't collect all of that.
P.S. Now marry that huge dataset with services that Accenture provides, among others:
"In February 2025, Vice News spoke to a former Accenture employee under the condition of anonymity. His project on the WhatsApp team for Meta required him to sift through images and decide whether or not they depicted child sexual abuse, which he coped with "through a lot of substance abuse". The former employee claimed to have witnessed multiple missed opportunities to protect children, and alleged that one colleague had previously been arrested for possessing child abuse materials. In a statement, Accenture said they are "committed to helping companies keep their platforms safe through services such as content, advertising, and compliance reviews."¹
¹: https://en.wikipedia.org/wiki/Accenture
Hard to see how that is remotely worth $1bn.
Fast.com has existed for 15 years yet isn't nearly as popular. It's easy to build a new speed test, but much harder to get people to use it.
Downdetector wins because of SEO. Most people don't get there directly, they google for "is $x down" and then get sent to downdecetor. Which from my understanding works by simply showing you how many people came to their site with those search terms. They don't actually check the sites.
Fast is a Netflix product so the fact that you've even heard of it is in direct relation to the weight of the brand that launched it.
speedtest.net has been the first search result on Google for "speed test" for decades. Partly the boost of domain SEO and partly the boost of it being an effective exit node for searches for that term for that long.
(Nobody searches "ookla" and nobody is going to search your tier-3 .com)
It takes more than money to supplant the name brand that every ISP games and every front line support worker uses by name
More than $1bn?? I don't buy it.
May I suggest a deep dive into to the concept of “time value of money”
selling peoples ip addres for some reason along with whatever privacy invasion tech they have.
There is something I never liked about Crapcenture - it's corporate culture is so weird and almost cult-like. There is no doubt they are successful but I question whether that model should even exist in the first place.