> price hikes appear designed to retain whales while tossing back small fry
Yea. That's how Enterprise SaaS has always worked.
I'm always going to prioritize F1000 customers over a small business. And small businesses in turn will end up using MSPs.
I'm not going to change my entire product roadmap to retain a customer generating $20k ACV when I can use the same effort for a customer generating me 6-8 figures of ACV.
Everyone from Broadcom to AWS to Snowflake to Databricks does this.
You get what you contribute for (either in money or code base contributions) - be it an OSS product or a paid product.
Welcome to business - the whole point is to make money, not art.
Edit: What's with the downvotes. Plenty of people come on HN for startup advice.
This is literally what EVERY company ends up doing. If you cannot target strategic or enterprise accounts, no one will give you money, and your product will inevitably be swallowed by a larger player.
And as employees it's critical to recognize this fact - this is the reason why your salary is getting paid.
I still don't get this from a strategy point of view. Let's say you have 100 "whales" who generate half your revenue and another 100,000 "minnows" who generate the other half. That's still half your revenue. As long as you are making a profit on the minnows, even if small, surely you'd want to keep them? I assume they must be profitable because otherwise third parties wouldn't be trying to poach them, to extend the fishing analogy a bit further.
(That's not counting the huge negative reputational effect that this change is having on VMware which is surely worth something.)
And when they'll become whales they'll call Broadcom since scrappy start-up aren't gonna keep acting like scrappy start-ups when they reach whale scale.
That's how major semiconductor companies operate. They'll only talk to you if you're a whale promising to buy volumes worth millions, they're not gonna waste time talking to hobbyists on the off chance that one of them might become a whale because the chance of a hobbyist building the next Apple in his garage is close to zero, and if by some miracle he does reach Apple scale, he'll still inevitably end up needing to call Broadcom later as the company expands and diversifies.
Try calling Broadcom or NXP asking for a detailed datasheet for a product that's not aimed at hobbyists, see where that gets you.
No one is going to invest into VMware stack, and if they become "whales" they will become that while running a different stack, and why should they invest into replacing it wholesale with customer-hostile Broadcom?
Larger companies can bundle more products so you get 4-5 things for the price of 1, plus you get dedicated support staff and account management.
Doesn't matter if you started off as a scrappy startup doing janky platform work - you inevitably migrate to the bundle because pricing is better for enterprises, and account management is much more streamlined.
They will, because someone on the board will know someone from vmwares board.
Besides, I've seen people buy oracle because all the big boys need Oracle, and we're a big boy now. Does anyone know what this 'database' thingy is? We really need one in this company, maybe it can store a list of all of our postgreses, mongodbs and mysqls? For management purposes?
> I've seen people buy oracle because all the big boys need Oracle
DevEx does matter. Just because your large doesn't mean you can rest on your laurels.
30 years ago IBM DB2 was the goto DBMS, before they got upstarted by Oracle DB2, and the Oracle DB2 customers began moving away to cloud hosted DBs such as AWS DynamoDB or Open Core offerings.
That’s not what happened with Google office vs Microsoft office. Microsoft is able to work office into whatever large company they want for the most part.
With VMware all the sales team needs to do is point to all their largest customers and say “we support these guys, are you sure your tiny company will meet your needs now that you’re a whale? We even have this convenient migration tool”. And the CIO won’t say no.
I'd argue they explicitly do the opposite. And that's why they could take back clients who went with alternatives - because they have a strategy where they can offer something from small office to multinational corporation, and have tailored packages and strategies at all levels.
And the problem with VMware is that to reach the scale where they are going to play ball, you're probably going to have the parts integrated already. Not a transition from small scale stuff where VMware integrated platform used to turn heads as migration path. Hell, at the scale Broadcom wants to only play there's no migration tooling available - and most is being built now to migrate away.
When you reach the scale Broadcom wants to focus on, you're going to have significant infrastructure spend already - and the CIO might get asked heavy questions regarding migration cost and time.
Frankly speaking, Broadcom strategy is one that does not admit new clients.
> Let's say you have 100 "whales" who generate half your revenue and another 100,000 "minnows" who generate the other half
When I was a PM in the Enteprise SaaS space I owned a product line that generated 9 figures a year in revenue.
Around 70-80% of that revenue came from around 15-20% of our customers.
Just 15-20 customers spending 7 figures year on my product could have allowed my PL to hit ACV metrics comparable to multi-billion IPOs, and we had WAAAAAY more high 6 figure and 7 figure customers than 15-20.
I cannot be bothered by an IT Admin at a local real estate broker's roadmap - I only had 10-30 engineers and we had way larger customers who generated us more money and required the same amount of effort.
Also, screwing over the minnows who are willing to spend some money is a huge gift to the competition. If another product grabs a large portion of the market, and now has more resources to innovate with, the whales may even start to switch over from the now mostly stagnant and overpriced option. I love(d) VMWare, but going forward it will have no place in any environment I control.
When you buy Enterprise SaaS products (DB, Cybersecurity, Cloud, Networking, Infra) these are backbone products.
If you can't afford to spend $300k a year on a product, you can't afford to hire the in-house staff needed to manage an OSS or internally built product.
There's a reason why Project Borg/K8s was invented by Google, not your local MSP, and why PostgreSQL's largest sponsors are AWS, Fujitsu, Google, Microsoft, NTT, and Broadcom.
The thing is = Broadcom is cutting off the MSPs, too.
Compare it with MS strategy since pretty much 1990 - they not only had offerings centered at small business, and medium business - they also didn't drop them once they truly got into the big league with Windows 2000 and Active Directory (yes, that was very much a watershed moment).
It can be a bit complicated to buy at the lowest end, but it's also been unspoken policy of them to somewhat tolerate software piracy so long as it didn't really impact sales - as well as offering great deals for people who could then get into track to get them more sales over time.
Some companies will cludge together things in house. Some people will learn and make MSPs and VARs. Some companies will become high value clients with white glove treatment from MS directly or designated support partners.
But they always made it so that they could get themselves an in without having to fight a big honking procurement contract to land a "whale".
As for 300k USD a year - outside of Silly Valley, it can be enough money to get custom linux distro developed and supported. Hell, with 300k USD/year, I could get a team to build and support a replacement for all core parts of VMware (compute, storage, networking). And if we didn't need to replicate same level because we were a smaller client and didn't need hyperconverged setup, we could do it for less.
> As for 300k USD a year - outside of Silly Valley, it can be enough money to get custom linux distro developed and supported
Erm, not really. One of my portfolio firms landed a multiyear runtime security contract in the Midwest around that much. And it was a mid-market prospect.
If a runtime security contract can land for that alone, including comprehensive Linux distro support will break 7 figures (at least based on the RHEL quote the customer has)
> it's also been unspoken policy of them to somewhat tolerate software piracy so long as it didn't really impact sales - as well as offering great deals for people who could then get into track to get them more sales over time
There's a reason why every vendor is moving to a SaaS and metric driven workflow. We can attribute which customers are break their contracts and gladly extracting a pound of their flesh for breaking the contract (as they would with us)
Even MSPs are consolidating, and there's a reason vendors like working with a Presidio type firm.
> If a runtime security contract can land for that alone, including comprehensive Linux distro support will break 7 figures (at least based on the RHEL quote the customer has)
The world does not end at US borders. Not everyone can exploit that, but it does not mean there's no market to be exploited.
> There's a reason why every vendor is moving to a SaaS and metric driven workflow. We can attribute which customers are break their contracts and gladly extracting a pound of their flesh for breaking the contract (as they would with us)
Microsoft might not have had the metrics originally, but to this day they play the game of getting people hooked on using their products licensed or not, and for all I heard or experienced they never went for inane extraction of pounds of flesh like Oracle does - which I suspect is because someone there might understand how it results in priority 1 becoming "how do we get rid of Oracle products", especially if the lock-in isn't too deep.
> Even MSPs are consolidating, and there's a reason vendors like working with a Presidio type firm.
A big issue with Broadcom strategy is that they are cutting off significant chunk of MSPs as well and preventing a healthy ecosystem of those with how they are changing licensing.
I'd say it depends on the product and level of effort that goes into supporting either group but yes, I'd generally lean towards this as well but it can be very difficult for a business to maintain this depending on their stage.
The upside to having a healthy number of minnows is, assuming it's low effort to have them onboard in a mostly self-services manner, then they can help the business grow in a healthy way while the team focuses on the larger and more complicated opportunities that may easily lead to nothing. If all you have are the whales it becomes feast or famine but if you have a healthy combination of small and large you'll be well covered. And of course you're banking on a percentage of the smaller customers eventually growing too.
Again, it's about the product and selling motion. Back to the topic, in this instance Broadcom did the math and they simply don't have a product that lends itself well to smaller shops and just isn't worth their time and effort to pursue.
Half of the revenue does not mean half of the profit. Spending any time working with minnows may be time not worked for big companies. With big customers you can have much higher profit margins when they want to buy expensive services.
To extend your analogy further. When you measure the return on investment for minnows vs whales, you realize that your time is worth more when you focus on the whales. So you end up diverting 100% resources to the whale operation out of your fiduciary duty towards your shareholders (you are a business at the end of the day).
Were this a private operation, they may be more leeway in terms of striking that balance between profitability and social responsibility. A good example of that is Patagonia (the outdoor apparel company).
If I even had an inkling that my portfolio startups in the Infra or Cybersecurity were not targeting large strategic customers who can subsidize their roadmap, I'd literally fire the founding team.
Not true. Minnows don’t even have a sales cycle for most of them, they will download and then pay for support via credit card and then support over email. New enterprise clients take 18+ months to land and then an entire account management team.
Minnows will still file support tickets and bugs, and clog up our support and engineering team's time, and then complain when we ignore them because we'd rather work on the support tickets that customer X who's spending $200k a year instead of you who is paying $1k a year.
I worked at a SaaS company that did a fairly significant price hike that priced out the smaller guys.
They did the math beforehand: something like 90% of our support budget went to companies that were contributing 5% of our revenue. Why not focus on the larger contracts that return much higher margins, and eliminate that high support burden that came with the smaller contracts?
From what I’ve read, this has been Google’s operating model. They will happily shut down a profitable business if it isn’t hundreds of millions of dollars worth within some time-frame. It’s a shame. I wish they would just spin them off instead.
I agree. If you did so much innovation already, but it doesn't generate the margins needed, a spinoff with minority ownership would make so much more sense and gain the ability to grow.
A lot of enterprise companies actually do this, but I have absolutely no idea why Google does not. Maybe some sort of antitrust paranoia?
If someone from GOOG can give visibility into the strat I'd honestly love it.
This is only a partially educated guess from a worm's eye view of FAANGs but if the product is designed to run on a FAANG's internal infrastructure, relies on their their internal-only libraries and development processes, etc., it would likely be more effort and risk to undertake the engineering to transform it into a standalone spin off business than it would recoup. Not to mention that key project staff would likely rather stay with the FAANG and initiate internal transfers rather than be spun off, further complicating the process.
I've never been at Google, but I've been at Yahoo and Facebook...
The problem with spinning off something that was designed and built as a Google project is it would take a lot of work to run it outside of Google.
What I saw in Yahoo wasn't too conjoined, ads, user login, and link tracking would be effort to move, but most of say Yahoo Travel was just regular PHP with regular MySQL underneath. You'd need a new packaging/remote installation tool as well, but ok.
Otoh, at FB and I gather at Google as well, everything is proprietary tools. Some of which end up sort of open sourced. But it's not like you can take a source dump and spin it up in another cloud; it would probably be a 6 - 12 month slog, and then you've lost whatever thin market you had.
Because Snowflake is spending 9 figures subsidizing you.
If you want a feature change in s3, if you aren't paying 8-9 figures, you ain't getting that feature change.
For example, if you want to S3 buckets to encrypt using Tiger hashing (for god knows what reason) if you ain't spending 8-9 figures, it ain't happening.
Some people do not get how important the long tail is for sustaining a tech platform business. Ultimately technology is interchangeable but attention, mindshare and network synergies critically depend on the people volume of the long tail.
They really have kicked their grass roots supporters to the curb with this change. They already used distribution/aggregation, it really wasn’t a huge lift for them to continue that chain.
Let's say I'm a mid-market business that doesn't want to roll out internally and I don't want to shell out for Broadcom (eg. Maybe I compete with their chips division).
I still want a throat to strangle, and Proxmox, while a solid product, doesn't have the headcount needed to support me.
A larger firm like Veeam can provide both Proxmox support and additional management support.
This way I don't have to use Broadcom, Citrix, etc.
The expected margins, pricing among peers, and GTM is entirely different from the Enterprise SaaS space.
It's the kind of product that is priced to be within an EM's or PM's discretionary budget for productivity tools, or out of pocket.
The strategy used to build a Basecamp and Notion is entirely different from the strategy needed to build a Kong or Noname, as the buyers are entirely different.
53 comments
[ 5.6 ms ] story [ 110 ms ] threadYea. That's how Enterprise SaaS has always worked.
I'm always going to prioritize F1000 customers over a small business. And small businesses in turn will end up using MSPs.
I'm not going to change my entire product roadmap to retain a customer generating $20k ACV when I can use the same effort for a customer generating me 6-8 figures of ACV.
Everyone from Broadcom to AWS to Snowflake to Databricks does this.
You get what you contribute for (either in money or code base contributions) - be it an OSS product or a paid product.
Welcome to business - the whole point is to make money, not art.
Edit: What's with the downvotes. Plenty of people come on HN for startup advice.
This is literally what EVERY company ends up doing. If you cannot target strategic or enterprise accounts, no one will give you money, and your product will inevitably be swallowed by a larger player.
And as employees it's critical to recognize this fact - this is the reason why your salary is getting paid.
(That's not counting the huge negative reputational effect that this change is having on VMware which is surely worth something.)
That's how major semiconductor companies operate. They'll only talk to you if you're a whale promising to buy volumes worth millions, they're not gonna waste time talking to hobbyists on the off chance that one of them might become a whale because the chance of a hobbyist building the next Apple in his garage is close to zero, and if by some miracle he does reach Apple scale, he'll still inevitably end up needing to call Broadcom later as the company expands and diversifies.
Try calling Broadcom or NXP asking for a detailed datasheet for a product that's not aimed at hobbyists, see where that gets you.
No one is going to invest into VMware stack, and if they become "whales" they will become that while running a different stack, and why should they invest into replacing it wholesale with customer-hostile Broadcom?
Larger companies can bundle more products so you get 4-5 things for the price of 1, plus you get dedicated support staff and account management.
Doesn't matter if you started off as a scrappy startup doing janky platform work - you inevitably migrate to the bundle because pricing is better for enterprises, and account management is much more streamlined.
Besides, I've seen people buy oracle because all the big boys need Oracle, and we're a big boy now. Does anyone know what this 'database' thingy is? We really need one in this company, maybe it can store a list of all of our postgreses, mongodbs and mysqls? For management purposes?
DevEx does matter. Just because your large doesn't mean you can rest on your laurels.
30 years ago IBM DB2 was the goto DBMS, before they got upstarted by Oracle DB2, and the Oracle DB2 customers began moving away to cloud hosted DBs such as AWS DynamoDB or Open Core offerings.
With VMware all the sales team needs to do is point to all their largest customers and say “we support these guys, are you sure your tiny company will meet your needs now that you’re a whale? We even have this convenient migration tool”. And the CIO won’t say no.
I'd argue they explicitly do the opposite. And that's why they could take back clients who went with alternatives - because they have a strategy where they can offer something from small office to multinational corporation, and have tailored packages and strategies at all levels.
And the problem with VMware is that to reach the scale where they are going to play ball, you're probably going to have the parts integrated already. Not a transition from small scale stuff where VMware integrated platform used to turn heads as migration path. Hell, at the scale Broadcom wants to only play there's no migration tooling available - and most is being built now to migrate away.
When you reach the scale Broadcom wants to focus on, you're going to have significant infrastructure spend already - and the CIO might get asked heavy questions regarding migration cost and time.
Frankly speaking, Broadcom strategy is one that does not admit new clients.
When I was a PM in the Enteprise SaaS space I owned a product line that generated 9 figures a year in revenue.
Around 70-80% of that revenue came from around 15-20% of our customers.
Just 15-20 customers spending 7 figures year on my product could have allowed my PL to hit ACV metrics comparable to multi-billion IPOs, and we had WAAAAAY more high 6 figure and 7 figure customers than 15-20.
I cannot be bothered by an IT Admin at a local real estate broker's roadmap - I only had 10-30 engineers and we had way larger customers who generated us more money and required the same amount of effort.
However, there's no end to the continous requirement of revenue growth. Isn't that why certain products are hated and how they eventually decline.
The duration and timeline may change depending on the product but it is most likely the eventual outcome of this strategy.
Is this factored in into the pricing or is there a plan to deal with it if/when that happens.
It comes down to roadmap.
Inevitably, if you've successfully managed to 0-to-1 a very successful product, you have the itch to start something new and actually own it.
This means the visionaries leave and build disruptors, while the remaining people don't care about the mission.
A large enterprise company can continue to make cutting edge innovations if they continue to incentivize that within their corporate culture.
> Is this factored in into the pricing or is there a plan to deal with it if/when that happens
Yes. Product owners work closely with sales to identify prospects who's roadmap and vision matches closely with your internal roadmap and vision.
If it's a large enough customer, they will always ask for a 6-12 month roadmap.
The issue is the competition does the same thing.
When you buy Enterprise SaaS products (DB, Cybersecurity, Cloud, Networking, Infra) these are backbone products.
If you can't afford to spend $300k a year on a product, you can't afford to hire the in-house staff needed to manage an OSS or internally built product.
There's a reason why Project Borg/K8s was invented by Google, not your local MSP, and why PostgreSQL's largest sponsors are AWS, Fujitsu, Google, Microsoft, NTT, and Broadcom.
And this is why minnows go to MSPs.
Compare it with MS strategy since pretty much 1990 - they not only had offerings centered at small business, and medium business - they also didn't drop them once they truly got into the big league with Windows 2000 and Active Directory (yes, that was very much a watershed moment).
It can be a bit complicated to buy at the lowest end, but it's also been unspoken policy of them to somewhat tolerate software piracy so long as it didn't really impact sales - as well as offering great deals for people who could then get into track to get them more sales over time.
Some companies will cludge together things in house. Some people will learn and make MSPs and VARs. Some companies will become high value clients with white glove treatment from MS directly or designated support partners.
But they always made it so that they could get themselves an in without having to fight a big honking procurement contract to land a "whale".
As for 300k USD a year - outside of Silly Valley, it can be enough money to get custom linux distro developed and supported. Hell, with 300k USD/year, I could get a team to build and support a replacement for all core parts of VMware (compute, storage, networking). And if we didn't need to replicate same level because we were a smaller client and didn't need hyperconverged setup, we could do it for less.
Erm, not really. One of my portfolio firms landed a multiyear runtime security contract in the Midwest around that much. And it was a mid-market prospect.
If a runtime security contract can land for that alone, including comprehensive Linux distro support will break 7 figures (at least based on the RHEL quote the customer has)
> it's also been unspoken policy of them to somewhat tolerate software piracy so long as it didn't really impact sales - as well as offering great deals for people who could then get into track to get them more sales over time
There's a reason why every vendor is moving to a SaaS and metric driven workflow. We can attribute which customers are break their contracts and gladly extracting a pound of their flesh for breaking the contract (as they would with us)
Even MSPs are consolidating, and there's a reason vendors like working with a Presidio type firm.
The world does not end at US borders. Not everyone can exploit that, but it does not mean there's no market to be exploited.
> There's a reason why every vendor is moving to a SaaS and metric driven workflow. We can attribute which customers are break their contracts and gladly extracting a pound of their flesh for breaking the contract (as they would with us)
Microsoft might not have had the metrics originally, but to this day they play the game of getting people hooked on using their products licensed or not, and for all I heard or experienced they never went for inane extraction of pounds of flesh like Oracle does - which I suspect is because someone there might understand how it results in priority 1 becoming "how do we get rid of Oracle products", especially if the lock-in isn't too deep.
> Even MSPs are consolidating, and there's a reason vendors like working with a Presidio type firm.
A big issue with Broadcom strategy is that they are cutting off significant chunk of MSPs as well and preventing a healthy ecosystem of those with how they are changing licensing.
The upside to having a healthy number of minnows is, assuming it's low effort to have them onboard in a mostly self-services manner, then they can help the business grow in a healthy way while the team focuses on the larger and more complicated opportunities that may easily lead to nothing. If all you have are the whales it becomes feast or famine but if you have a healthy combination of small and large you'll be well covered. And of course you're banking on a percentage of the smaller customers eventually growing too.
Again, it's about the product and selling motion. Back to the topic, in this instance Broadcom did the math and they simply don't have a product that lends itself well to smaller shops and just isn't worth their time and effort to pursue.
Comparative Advantage https://en.wikipedia.org/wiki/Comparative_advantage
Were this a private operation, they may be more leeway in terms of striking that balance between profitability and social responsibility. A good example of that is Patagonia (the outdoor apparel company).
Privates and startups do the same thing.
If I even had an inkling that my portfolio startups in the Infra or Cybersecurity were not targeting large strategic customers who can subsidize their roadmap, I'd literally fire the founding team.
They did the math beforehand: something like 90% of our support budget went to companies that were contributing 5% of our revenue. Why not focus on the larger contracts that return much higher margins, and eliminate that high support burden that came with the smaller contracts?
I agree. If you did so much innovation already, but it doesn't generate the margins needed, a spinoff with minority ownership would make so much more sense and gain the ability to grow.
A lot of enterprise companies actually do this, but I have absolutely no idea why Google does not. Maybe some sort of antitrust paranoia?
If someone from GOOG can give visibility into the strat I'd honestly love it.
The problem with spinning off something that was designed and built as a Google project is it would take a lot of work to run it outside of Google.
What I saw in Yahoo wasn't too conjoined, ads, user login, and link tracking would be effort to move, but most of say Yahoo Travel was just regular PHP with regular MySQL underneath. You'd need a new packaging/remote installation tool as well, but ok.
Otoh, at FB and I gather at Google as well, everything is proprietary tools. Some of which end up sort of open sourced. But it's not like you can take a source dump and spin it up in another cloud; it would probably be a 6 - 12 month slog, and then you've lost whatever thin market you had.
A Snowflake will always be treated superior to Joe with his personal project.
Snowflake's product suggestions will always make AWS's roadmap, but Joe's will not.
Snowflake will (most likely) get a 70-90% discount per object while Joe will be paying a 100-300% oversell.
If you want a feature change in s3, if you aren't paying 8-9 figures, you ain't getting that feature change.
For example, if you want to S3 buckets to encrypt using Tiger hashing (for god knows what reason) if you ain't spending 8-9 figures, it ain't happening.
Customers are not equal.
Small little guys managing their own little clouds in their home eventually become CTOs. Vmware is going to learn an important lesson here.
Oracle is generally and universally hated. Welcome to that club, Vmware.
If you want to use Proxmox but also need support, you'll end up paying Veeam, which itself uses the same GTM model I mentioned.
I still want a throat to strangle, and Proxmox, while a solid product, doesn't have the headcount needed to support me.
A larger firm like Veeam can provide both Proxmox support and additional management support.
This way I don't have to use Broadcom, Citrix, etc.
https://basecamp.com/small
https://www.inc.com/magazine/201606/jason-fried/saying-no-to...
https://medium.com/signal-v-noise/bigger-prices-bigger-probl...
The expected margins, pricing among peers, and GTM is entirely different from the Enterprise SaaS space.
It's the kind of product that is priced to be within an EM's or PM's discretionary budget for productivity tools, or out of pocket.
The strategy used to build a Basecamp and Notion is entirely different from the strategy needed to build a Kong or Noname, as the buyers are entirely different.