You don't want to sell to everyone - you need to make sure the cost of retention+support is significantly less than what they are paying.
If a customer is too troublesome or spending too little (for a company like VMWare, a $8m contract is on the smaller end) this is the politest way to fire them.
It sounds like this customer is a legal firm, which makes sense - they tend to have a much smaller IT footprint.
Every dollar isn't the same.
Edit: to elaborate - it's all about margins. Is that $8m worth it if in aggregate you spending an equal amount of time on that account as you are on a $20-30m account? It ain't
I work at a company that got sold by another company with an indefinite LTA (long term agreement) in place. Former owner did not honor the LTA, but at least was transparant and timely about it. We got two years to sever several key processes. We will end up severing all ties, since the business risk of leaving open ends without the LTA is too big for my risk appetite. Fool me once. And that's our business consequence of not honoring an LTA even when the former owner severed the ties in good faith. If the buy-in is not there for both parties to the contract, then there is no ongoing relationship.
Chances are this is not actually surprising for this customer. There's probably been a lot of talk between the sides, a lot of expectations that weren't met both ways.
What you're describing at your own company seems to be that same thing, no? Relationship wasn't going well, and it's gotta end somehow.
Now that's not to say this should be normalized, it shouldn't. But there needs to be some way to say "hmm, it doesn't make sense for us, business wise"
consider that, for shitty, demanding customers, the 'fair price' may be 12x more.
$8M probably doesn't justify the effort, and they may well use $10M+ in support, resources, or potential liability.
Good faith in the contract is expecting to give something in the contract and then meaning it. They will give good support -- for 100M.
The strategy you described has no upside for the vendor (in this case Broadcom VMWare).
Absolutely.
For an Enterprise SaaS generate $13.5B in revenue in FY23, you can realistically assume mean deal size is mid-upper 8 or low 9 figures.
Everyone and I mean everyone is using them
Tech is littered with the twittering ghosts of companies that had something "essential" and "irreplaceable". The strategy is "a retreat to the high end" and a self-serving re-definition of "who are customers are" and "what we do".
As smaller shops switch away (the Proxmox folks must be in a pretty good mood), the job market will be full of people who know the alternatives, but not VMWare products. And when "the real thing" is so incredibly expensive, why not switch to something comparable that your people already know?
When you reach VMWare size ($13.5 BILLION in revenue in FY 23), $8M is chump change.
The margins on support might not make sense.
With the level of revenue VMWare had, it's realistic to assume mean deal size is in the 8-9 digit range.
> what retention
TAMs, CSMs, constant PFRs, constant escalations, etc
> This is basically a polite way to fire a customer.
I'll throw out that raising the price like this is also a strategy to increase profit. In the absence of a source for your claim, I'll go with that instead.
> Is that $8m worth it if in aggregate you spending an equal amount of time on that account as you are on a $20-30m account?
If that's how these decisions are made, the person making them needs to be fired, and fast. They should have considered that they could handle both accounts. Increasing the number of customers is called expansion.
I do this for a living, first as a PM and now as a VC. These multiples are fairly common knowledge for anyone in Enterprise SaaS, and have have worked with accounts teams that have done this and have recommended similar strategies (at a smaller scale) to portfolio companies.
> the person making them needs to be fired, and fast. They should have considered that they could handle both accounts. Increasing the number of customers is called expansion.
I can expand the # of customers by selling 1,000 $1 contracts - that doesn't materially impact my company's financials.
You want to increase your (Revenue-COGS).
If you need to drop a needy customer to increase COGS because you have mistiered accounts caused by an account team targeting a hefty quarterly bonus (happens everywhere, and caused a big shitshow at a major cybersecurity startup recently that brought a federal investigation) so be it.
I think it depends? Let's assume that both accounts are profitable overall, but one is massively more profitable. The logic to cut off the $8m account makes sense if you can move those resources towards more $20m accounts. But that isn't something that is necessarily possible, is it? There's only so many of those $20m customers out there.
This obviously changes if the $8m customer is not profitable or is just _that_ unpleasant to work with.
At VMware size it absolutely is.
I've worked on accounts at similarly sized vendors and even AV or Endpoint spend will be in the $1-5M range alone even if per unit price is in the $20-30 range.
With a backbone product like VMware, you can safely assume spend is 4-6x as you will be bundling network security along with IT infra.
If you spend 75 cents to earn a dollar on one customer and can spend 25 cents to earn a dollar on another customer, you'd be dumb not to take customer 2.
> Just emailed the CEO of Broadcom and got his OoO: “Thank you for your email. The exorbitant price increases are due to inflation and macroeconomic conditions. Also, the taxes on my mansion in Cabo are sky high!”
Not that VMWare customers aren't faced with challenging circumstances...
I can’t wait for the AWS sob stories.
But even if it wasn’t, aws could add a zero to the end of their prices next year and a lot of companies will not be able to migrate
When market conditions change significantly you can see massive degradation of the business somewhere.
VMware shake down could certainly affect the cloud market significantly.
Their strategy is to go after a few of the biggest organizations, and not a bunch of small mom-and-pop shops. Just like IBM with mainframes, which still seems to be profitable for them.
This will be a shift for on-prem deployments. For cloud solutions running VMware in AWS/Azure, unless those renewals go up significantly, most of VMwares mrr is already locked in and the people running these setups generally refuse to learn how the cloud works so it’s a pure lift and shift mentality.
But the pattern begins, broadcom buys VMware does nothing increase prices. Why would any sane client stay around to get VC squeezed.
Also, what are the alternatives proprietary and open? And what are those solutions missing? Hopefully this is good news for the open project, getting more commercial support and more users.
Also sounds like good news for people like Oxide. If costumers are already considering leaving VMWare and need to do a lot of software changes, they might be willing to make a larger step for new hardware as well.
Beyond VMs they attempted to branch out into containers with a version of Kubernetes (VMware Tanzu) which I believe didn't get many sales.
The competition are roughly:
- All the cloud providers
- oVirt (formerly Red Hat's RHV, probably the closest thing to vSphere in terms of UI and deployment)
- Kubenetes + Kubevirt
- Openstack, probably
- Virtuozzo
- Virtual Box
- Proxmox
- Nutanix
- Single node installations using things like virt-install
- Rolling your own management on top of KVM or Xen (surprisingly popular but I wouldn't recommend it myself)
Algolia overage charges are a fortune. If we kept on going we’d be burning an equivalent software engineer worth of salary.
Now we give equivalent sum to a real human.
AV or Endpoint spend at a F1000 is around that much alone.
A backbone vendor like VMWare can and does charge 4-6x more ime.
They paid, didn't they?
And people enter into unfair contracts regularly without realizing it or understanding it.
So the answer is no, you don't have any information not found in the link, and everything you wrote was speculation.
> You want to increase your (Revenue-COGS).
Exactly, and without knowing anything about the financials of this deal, there's no way to know if your claims are correct. It might have been an attempt to take advantage of someone that built on their product, expecting that they'll pay more.
If you don't trust me, go chat with CROs, Product Leadership, or Execs are similarly sized companies.
I agree this is tribal knowledge, but that's how most pricing and renewal negotiatons occur, with maybe a bit of RevOps black magic to justify a decision one way or the other.
Care to elaborate?
I can say this is a company where the CRO "left".
It seems nobody can believe your assertion that a company would deliberately fire a profitable customer by quoting an x12 contract increase. There's no point asking me to talk to people in my business. You're going to have to point to publicly accessible resources.
Your SaaS vendor will hike prices or fire people until they get to this number. Why ? Because VC/Wall Street told them that if they want a Meta/Apple/Microsoft like valuation or a good upround they need to hit this key metric.
Over the next 2-3 years every SaaS that is priced by some sort of market(so not passion projects/ projects run by guys who don't look at comparables) will at least try to hit the above metric. Whether that is massively laying off people(ring any bells?) or massively hiking pricing depends on the product in consideration.
The era of "profitable customer" is over. Now it's all about how profitable he is, and 1.5 mio in revenue per employee is the benchmark.
This sounds like a mis-tiering of an upper market or strategic account by an account team.
That said, Margins and RPE are critical now.
A lot of this is also because there are too many damn SaaSes now, and enterprise products bill monthly now as well.
In 2013 the average was 8 SaaS products per company and in 2023 around 80 (I can't remember the blog link I got this from but it was a fellow VC or CRO who pointed this out).
There isn't as much money floating around anymore as everyone is at their limit, and it's getting ridiculous to spend a couple hundred thousand on some random SaaS that's unneccesary.
Sure that's not a UI and doesn't include a lot of other stuff that VMWare does, but it does illustrate what a gigantic amount of money 100 million $ actually.
The underlying open source tools are so good, if you have the money to have real engineering team and combine them in interesting way and to costumize them for your needs, 100 million $ seems like it would go a very, very long way.
Specially when we are talking 100 M$ every year or something like that.
Though I would argue that relying on cooperative/live migration in the first place is a mistake.
If you sign up to AWS on your own credit card, there was no inbound or outbound sales motion at all.
AWS's self service motion is extremely efficient for lower tier customers (as the upside is basically infinite as you basically spend $0 to generate N dollars). <- This is also why YC's Request for Startups is mentioning Open Core startups, as the open source offering acts as a loss leader (eg. Kong, Hashicorp, Pulumi eventually)
Traditionally, companies like VMWare didn't support such a model and your only way to purchase was to jump on a call with their account team, but companies are transitioning to that to target low tier sales.
VMware was bought at about 400k rev/employee, Broadcom was running at about 1.7mio per prior to that. Tan is essentially running a PE buyout shop that is listed, and he uses a lot of leverage. He might not get it to 1.5 mio but he wouldn't have bought it if he thought it could churn out anything less than like 1.2(over 6-7 years).
Pretty much.
> Tan is essentially running a PE buyout shop that is listed
I'd disagree on that. A PE play would severely underpay engineering and look for a short (2-3 year) turnaround by re-IPOing. Meanwhile, Broadcom tries to hold portfolio companies and pay AWS level salaries to hires.
Instead, Broadcom is using the same strategy that PANW and ZS are mimicking to own the F100 market.
There is severe vendor fatigue in the market, as various different apps are viewed as a massive cost inefficency and there is a very real chance of different apps becoming an attack vector (eg. Slack at Uber).
The only solution for this is vendor consolidation, as this
1. Limits the amount of RFPs you need to vet significantly
2. Narrow down any spending inefficiencies
3. Limits your attack vectors with a vendor breach or compliance overheads
This is why Niklesh Arora announced "Platformization" at PANW's earnings call a couple days ago, and is the same strategy Broadcom has been using for almost a decade, and what ZS+CRWD have started using as well.
This is a demand from F1000s now as well as institutional investors, and a massive reason why Broadcom's stock has gone from $40 to $1300 in a decade.
The upside is you don't scare off other customers when they hear reports of a 12x price increase.
Right now, in response to this post, VMWare's biggest customers will be drawing up plans for if they need to walk away at the next contract renewal. And they'll want those plans to be extremely credible - it's worth spending $5 million trialling and integrating a competing system, if it stops your $50m contract turning into a $150m contract.
Broadcom goes into these deals knowing it will hemorrhage a lot of customers and they don't care, they know they can make it up on those who are forced to stay. They don't care about reputation at all.
Usually that means prices increasing by ~10% every year, so that it's never quite worth it to switch to an alternative system.
"[Broadcom's methodology] was not based on taking existing products and raising their prices" Broadcom CEO Hock Tan Oct 26th 2022
It sounds like this was a (comparatively) PITA small-fry legal / law office.
Maybe I'm completely out-of-touch, but an 8M$ contract should be sufficient to get "appropriate deals"?
If this scares away customers spending less than that, good for them. It doesn't make a dent when a F1000 is spending tens of millions a year on VMware and purchase by committee and will get sweetheart deals by being strategic accounts.
It has. Product/brand trust matters... the most obvious example being Google, who have had an extremely hard time getting their stuff to be actually used as they have shown time and time again that even if you're the biggest game publishers on this rock, you won't be able to rely on Google keeping up their end of any deal.
At Enterprise SaaS level it does not.
You do not have an individual purchaser - the deal will be vetted by committee for 1-3 quarters.
> most obvious example being Google, who have had an extremely hard time getting their stuff to be actually used
Extremely hard disagree. GCP does well in enterprise/strategic accounts where a customer in some way competes with Amazon (which is around 20-30% of the F1000) and also doesn't have a significant Windows presence (minimizing the need for Azure+MS Support)
It should be:
(Contract_1_yearly+...+Contract_n_yearly)/num_contracts
Anyhow the customer signed the deal (https://twitter.com/cioontherun/status/1761882689886433742), implying they were underpaying and had the capacity to pay the true rate.
Also just to be clear I agree with your initial comment - there are lots of reasons even big customers can be worth repricing based on their true costs and strategic relevance. And at less than 0.1% of revenue VMWare can certainly afford to risk it.
If they are sufficiently locked in, this might have been the best option to migrate away without major disruption.
itsdrewmiller is right
I can guarantee for a fact that a F1000 will be spending 8-9 figures for a VMWare type product (I've worked on similar products as a PM).
$8m is low.
> Which is what people are upset about
All that matters is purchasing committees are ok (not happy, not sad - just meh).
All I see here is just IC engineers pouting.
Yeah, maybe. Without some real figures this is very speculative
> TAMs, CSMs, constant PFRs, constant escalations, etc
You seem to be assuming the customer takes up man-weeks or months of support and legal, not particularly typical for a repeat customer in that game - not sure how else you could account for the margin on $8m being completely eaten up
So... speculative and unconvincing
These kinds of numbers are closely guarded for competitive reasons.
Go chat with the sales or account teams at your employer - they can confirm in broad strokes what I've mentioned.
8m is 0.6% of their revenue, and they grew in revenue 4% from 2022 to 2023[1]. Losing 0.5% kinda hurts when they are growing that slowly. That 8m can't be that easily magicked out of existing customers or the freed up resources supporting that customer they fired. I know nothing about their business or financials however
[1]https://finance.yahoo.com/news/vmware-reports-fourth-quarter...
My hunch is an Account Team gave a heavily discounted deal to land an upper market customer in order to meet quarterly commitments or get a nice commission.
The customer ended up accepting the $100M price-point as well (https://twitter.com/cioontherun/status/1761882689886433742), so they were clearly mis-tiered. If you can afford (what I assume is) a multi-year contract with a 9 figure TCV, you are generating 9 or 10 figures in revenue.
I've heard of similar shenanigans down the grapevine at w/ Cisco and Dell EMC and absolutely wouldn't be surprised if a similar thing happened at VMWare.
I can understand VMWare doesn't want to keep doing this, but with $8M, assuming a 100k annual salary, you can dedicate 80 full time employees to this customer yearly.
I understand 8M is nothing next to 13.5B, but that can't be unprofitable, can it? Or the contract is too low.
1. "assuming a 100k annual salary" - the cost of salary is 1.75x stated wage, so in reality it's $175k
2. "you can dedicate 80 full time employees to this customer yearly" - you mean you need almost a hundred bodies per customers?!? In fact, this is a major reason to fire a customer. It means they are demanding and trying to outsource their entire org to you. Either pay up or piss off.
> that can't be unprofitable, can it
Nope.
> Or the contract is too low.
Exactly.
I know, this was ballpark thinking, it's not a fundamental flaw in my thinking. 100k with taxes is still a 60k salary and that's about what I was thinking. Not great everywhere, but still good in many locations. I also don't know how much support people are usually paid, and 200k is still 40 people.
Now, indeed, I wasn't imagining hundreds of people per customer required. I'm willing to believe it's realistic for some kinds of businesses, for very big customers.
8 million is quite close to the low end of 8 digits. And at the high end of 8 digits, your assumption woud already mean they have only a few hundred customers in total, which is about as far from "realistic" as can be.
That tends to be good enough for a PL within a larger organization, which VMWare is now.
Zscaler uses the same strategy and PANW is transitioning to as well.
Most of this is tribal knowledge you learn working a Field facing job.
Like I've said before, talk to SEs, AEs, and TAMs at your employer. They can confirm similar strategies and numbers.
When I worked with VMWare, it also wasn't uncommon for us to have support on site. If we were doing a big migration or had a significant issue, they'd send someone out for a week+.
They likely paid a premium to have support willing to travel for significant spans of time, plus now they're paying for hotels, daily stipends, etc.
Then you need labs for engineers to try to replicate problems, office space, general overhead like taxes and payroll providers, etc.
Then there's costs to the platform itself from some customers. E.g. small customers tend to be less homogenous. Every Fortune 500 has a SAN, many small companies don't. Now you need an "integrates with EMC" storage option for big companies, and an "integrates with everything else" option for small companies. Big companies want a service mesh, small companies want a billion different networking configurations, etc.
Basically big companies tend to resemble each other and view their differences as a competitive advantage they will pay their vendors to support so they can keep. Small customers diverge a lot, don't want to and/or can't afford to pay for the changes they need, and are generally more price-sensitive due to their ability and ability to switch vendors.
So simultaneously the small clients are profitable, but at a lower margin than bigger customers, and more willing to migrate away. Rather than taking a risk that they can do enough development to keep the small customers, they'd rather double down on the customers that can't switch, raise their margins, and cut their risk.
F1000s will spend the same amount on AV or endpoints alone.
For backbone infra, you can expect a price that is 4-6x higher, as network security, IT infra, etc is bundled.
Large enterprise purchasing committees are foremost concerned with risk, you absolute dolt.
The risk that price will jump 12x, requiring all kinds of scrambling, is fucking awful.
I don’t know what planet you’re from, but you gotta stop acting like it’s this one.
I worked at a F500 that used VMWare; they'd probably pay a 12x subscription fee. Not because they want to, but because they started trying to get rid of their mainframe 30 years ago and 70 something percent of their business still relied on the mainframe last I was there a decade ago.
VMWare will likely go the same. It'll take them a year or two to PoC an alternative and get contracts signed, another 5 to move the low hanging fruit, and a decade to finish off the long tail of stuff that won't migrate easily.
We're talking migrating stuff running versions of Java that were EoLed a decade ago, using an app container (like Tomcat) that was also EoLed a decade ago, and trying to get that all working on a new infra stack.
And it's not just one of them, there are thousands of apps, many of which are outdated or have bizarre vendor requirements you'd likely never accept (like "we will not support you if you run this bog-standard Django app on a VM").
Stuff like this moves slooooowwww, especially at companies who only use tech to enhance their primary business (as opposed to tech being their primary business).