ZS’s CEO will be on Cramer’s Mad Money tonight.
Rob
6 Likes
Give some, take some. At this point in time there is not a single serious competitor in the world doing what Zscaler is doing. If they do, please name it.
Not my field of expertise but what about this company: https://www.iboss.com/why-iboss-beats-zscaler/
I brought this up few months ago on this forum and since then, they seemed to have signed up some notable customers.
iboss might not be a ‘serious competitor’ considering they are small even compared to Zscaler (and Zscaler protection gets better the more customers they have – networking effect) but can’t they be described as ‘doing what Zscaler is doing’?
Zscaler is certainly ‘disruptive’ but it seems like their idea can be copied. And if they can copy it, why not Amazon who has the infrastructure than can go toe to toe with Zscaler? I only ask because Amazon has competes (or at least tries to) with everyone. Think Netflix, or Mongo or Elastic…etc. Granted Mongo has done a tremendous job of defending it position. And so will Elastic, I think.
So I think Zscaler’s Moat may not be it’s tech, but rather it’s first mover advantage. The bigger (more customer) it signs up, the better its service. The better its service…easier to land new customers…and so on.
Just sayin’.
TJ
Long ZS
8 Likes
https://www.iboss.com/customers/
Here is their customer list. They do feature highly Melody Ford somewhere in California.
Sears is a customer…errr, was at this point. They do have REI, which may be their marquee customer.
They are targeting small to midsize customers and the education market. Customers Zscaler does not serve except in branch office bids. I doubt very much Zscaler seems them at all, or rarely at best…just sayin.
Here is their Gartner review. Not review from just 8 months ago. Hardly an enterprise ready product just spinning its wheels wanting to take over your corporate network like Zscaler did 5+ years ago (in its entirety for GE) or a few years ago for Siemens. Again…just sayin…
Tinker
4 Likes
This assumes that 85% market penetration is achieved in five years
85%?? When you start with these kind of assumptions the rest of math doesn’t matter.
3 Likes
https://www.srgresearch.com/articles/switch-router-revenues-…
I’m not assuming 85% market share or the like. What I do know is Cisco still has more than 50% share in routers and switches and it is as about 15 points higher or so before Arista came along.
Latest estimate from Gartner from last year found that Zscaler has more than 55% market share in SWGs, although how much more they had no certainty.
The space is still relatively small and young. However, Zscaler can double its current customer base and still not hurt Palo Alto or Cisco or HP or whom ever who is dealing in appliances not much worse than what happened to Cisco with Arista.
The importance of this is that when a disruption happens like this the incumbent has no immediate existential reason to cannibalize their business to destroy the insurgent. By the time the insurgent gets to that point it is over.
Cisco still sells proprietary silicon and made a small effort to open its OS software, but nothing more against Arista.
The problem Arista had is that the market it was Disrupting was a nice niche but hardly the industry.
We shall see if SGW becomes more than a nice niche like Arista created, or more an industry like Cisco created and Palo Alto subsequently created.
In 2010 Palo Alto was smaller than Zscaler is now, was still private,
and it’s NGFW was still a growing niche so small that customers were still learning about it from referrals as most had no idea what NGFW was nor who Palo Alto was. At the time Cisco dominated the security market for data centers.
Certainly, those with common sense rule of thumb knew that once the market grew Cisco would just build its own NGFW product and stymie Palo Alto.
I mean serious de ja vu happening here. Whether it plays out the same, as it almost always does when a new platform that is of enough significance disrupts the old, we will have to see.
iBoss is years behind Zscaler. Their claim to fame is creating NGFW gateways in the cloud using containers. Their primary market is education and they are working to improve their enterprise offerings. If you are Henry County whose computers were down for a month last month (practically shut down the court system) iBoss may be a viable choice but Zscaler would have no interest in that customer unless an SI made the sale and did all the work.
Even then I’m not sure iBoss could handle the entire system.
iBoss does have a concise blog on why virtual machines spun in the cloud don’t work. The first reason is bandwidth cost is atrocious. I’m sure Palo Alto’s edge solution mitigates this to an extent, but it may also explain why we are hearing reviews as to how expensive it is.
This speculation aside, no one can doubt the repeated history of disruptive innovation. The pattern is well documented and has repeated itself over many different disruption cycles. Just happens that the Palo Alto success story practically mirrors,
word for word where
Zscaler currently stands. I mean just change the company names and Zacks could run the same stories
That were written 9 years ago, 6 years ago and so on until Palo Alto became the leading incumbent.
Zscaler has a much more disruptive product that creates a far more difficult innovator’s dilemma. Palo Alto, like Arista, did not eliminate the hardware model for pure software. Cisco, in neither case, had to scrap their business model and way of life. To do what Zscaler does you have to stop selling hardware, the replacement cycle for hardware, the annual maintenance contracts for the hardware, the software updates for the hardware, the consulting fees…and simply sell a SAAS service.
We shall see how effectively Palo Alto and Cisco deal with this. I honestly don’t know. I am just presenting the well known textbook framework as to how these things have historically played out time and time again. Exceptions of course.
Maybe iBoss dramatically improves its product and created another disruption, who knows.
Tinker
31 Likes
https://www.canalys.com/freereport/latest-free-report?id=146…
Here is what Palo Alto has done since 2016 in marketshare. They’ve upped Arista and may be the market leader by now as that is a 2018 chart.
This is what Palo Alto has done without Cisco suffering much or at all from an innovator’s dilemma.
Zscaler is bringing something obviously more disruptive and this risky as well. Will all cloud security become the technology standard as NGFW did? 9 years ago NGFW was a small, barely known niche.
I don’t know but that is the reason to invest in Zscaler. It may hit or it may miss. But it is truly a disruptive enterprise as (Zscaler’s CEO likes this analogy, and it actually is very analogous without being hyperbole) as Sales
Force disrupted the entire CRM market despite the incumbents, like Siebel who was dominant and then hit a Sales Force wall,
putting out hybrid cloud solutions for their customers to keep them from switching.
It may not work out that way on this market but it is easy to see how it might.
Tinker
10 Likes
This assumes that 85% market penetration is achieved in five years
85%?? When you start with these kind of assumptions the rest of math doesn’t matter.
The 85% is not pulled out of a hat, it’s an empirical number that’s been used people like Geoffrey Moore of The Gorilla Game fame. What’s the market PENETRATION of footwear? I would guess very close to 100%. BTW, the 85% does not apply to a single vendor but to the technology, the reason the table has five columns with increasing market SHARE by one supplier of the technology.
The five year figure IS pulled out of a hat! That’s the number that should have set off your alarms!
But there is a reason for it. SaaS/Cloud is a business model that allows for very fast implementation. Let me go back to my IBM sales rep days in 1963. You had to make the sale. Once the contract was signed IBM had to teach the client’s staff to program. They then had to write the software for their company. The client had to build the computer room, air conditioning, power supplies, false floors, etc. If everything went smoothly, nine months after the sales contact was signed the computer was delivered. Then the client had to be nursed through the implementation. Now the client already has PCs and trained operators, all they need to do is connect to the suppliers URL (OK, so I skipped a few details). The point is that cost and time to implement has shrunk very much like the size of transistors has shrunk. Today many systems have over 10 billion transistors. Had I proposed a system with a million transistors in 1963 they would have carted me off to the loony bin! Yet I would have been short by FOUR ORDERS OF MAGNITUDE.
Ray Kurzwei said that not only are things happening ever faster but that time itself was accelerating. My point is that SaaS/Cloud is now so easy to implement based not only on the product/service offered but also on the in-house technological expertise at the customer, that the “S” curves are going to become very compressed in time. The 5 years to 85% market penetration implies that the technology has a lifetime of around 15 years based on old empirical data. I believe that the three equal time durations of the “S” curve will change making the initial take up and the later cash-cow segments longer and the middle fast growth will be steeper and shorter. Using that model, the 5 years to 85% market penetration gives the technology has a lifetime of around 25 years. How long do you think Saul’s SaaS picks will be around? Five years? Ten years? Disruption will happen ever faster.
Denny Schlesinger
Transistor count https://en.wikipedia.org/wiki/Transistor_count
49 Likes
Sales Force disrupted the entire CRM market despite the incumbents
Yest Salesfore only has 20% market share in CRM. Still Siebel is used by all major US airlines companies for loyalty program, same for major consumer product companies, etc and the install base of Siebel more or less survived, CRM took lion share of the new implementation but was not able to dislodge the install base.
Of course, that doesn’t mean ZS cannot be successful. But, even in a wildly successful stories like CRM, the data shows the market share is much less than the mind share.
1 Like
The 85% is not pulled out of a hat, it’s an empirical number that’s been used people like Geoffrey Moore of The Gorilla Game fame.
Digital camera today replaced films, 100%. But digital camera was first invented way back in early 80’s and when digital picked up in 2001, still it took 10 years for the films to die or digital to take market share over 85%, just saying.
Now, ZS is not bringing such a complete paradigm shift and it needs to be seen whether they can sustain 50% market share within the non-appliance space.
Digital camera today replaced films, 100%. But digital camera was first invented way back in early 80’s and when digital picked up in 2001, still it took 10 years for the films to die or digital to take market share over 85%, just saying.
I’m afraid you are missing the context of my post which was laid out in plain text:
Looked at in “S” curve terms – very much in back of the envelope terms. The table below shows why SaaS stocks have seemingly absurd valuations, a compounding of account growth and per account revenue.
While I took the opportunity of the ZS Analyst Day thread to post it, it’s not ZS specific. Let me explain where the 85% number comes from. Growth in nature, from yeast to technologies, has a growth rate history best described by the sigmoid of “S” curve.
https://www.google.com/search?q=sigmoid+of+%22S%22+curve.&am…
S-Shaped Growth Curve
S-shaped growth curve (sigmoid growth curve) A pattern of growth in which, in a new environment, the population density of an organism increases slowly initially, in a positive acceleration phase; then increases rapidly approaching an exponential growth rate as in the J-shaped curve; but then declines in a negative acceleration phase until at zero growth rate the population stabilizes. This slowing of the rate of growth reflects increasing environmental resistance which becomes proportionately more important at higher population densities. This type of population growth is termed ‘densitydependent’ since growth rate depends on the numbers present in the population. The point of stabilization, or zero growth rate, is termed the ‘saturation value’ (symbolized by K) or ‘carrying capacity’ of the environment for that organism. K represents the point at which the upward curve begins to level, produced when changing population numbers are plotted over time. It is usually summarized mathematically by the logistic equation.
https://www.encyclopedia.com/earth-and-environment/ecology-a…
I highlighted the parts that are important for our discussion:
-
The “J-shaped curve” is when growth kicks into high gear forming the bottom of the "S curve. Empirical evidence as reported by Geoffrey Moore happens around the time the technology – NOT ANY ONE COMPANY SELLING THE TECHNOLOGY – approaches 15% market penetration.
-
“K” is where the top of the “S” is put in place. Empirical evidence as reported by Geoffrey Moore happens around the time the technology – NOT ANY ONE COMPANY SELLING THE TECHNOLOGY – approaches 85% market penetration.
That’s where the 85% comes from. The “S” curve is a generic model of growth rates, it can be slower or faster depending on the organism or technology it is modeling. Geoffrey Moore is a Silicon Valley marketing consultant and his experience when he wrote his marketing books Crossing the Chasm and Inside the Tornado was mainly in Silicon Valley high tech up to the late 1990s. The SaaS business model was yet to come. I make this point because until the appearance of the SaaS business model technology “S” curves could be divided into three equal length periods:
1- Slow growth from inception to 15% market penetration. Many technologies die in The Chasm, the subject of Moore’s book Crossing the Chasm.
2- Those that survive enter a fast growth phase. The start of that phase is the subject of Moore’s second book Inside the Tornado where 100% growth happens coming off a small base. In The Gorilla Game Moore advised to “Buy the basket and sell the losers.” What basket? The companies selling the technology because there was no way of telling which one would come out the winner! I came to the conclusion that this was not the best investing advice, I prefer to wait to see who the winner is, in other words, buy the winner post or late Tornado.
3- The third and final period starts as market penetration approaches 85% (there it is again) when the technology has saturated the market making fast growth impossible.
The most powerful innovation of the SaaS business model is EASE of IMPLEMENTATION. Complexity is a killer. In our SaaS companies, longer sales cycles are one example of complexity. The need for a large sales force is another. Anthropologist Jared Diamond says the “opportunity” not “necessity” is the mother of invention. In high tech Jeff Bezos is a prime example, he told how having the infrastructure in place (internet, credit cards, PCs, delivery, etc.) was what made Amazon possible. SaaS was made possible by the knowhow at the clients and by cloud delivery. One important consequence of the ease of implementation is that the middle period of the “S” curve is compressed in time, adoption rates – growth rates – are higher but the time to reach market saturation is shorter. This in no way affects the duration of the initial and final segments of the “S” curve.
The takeaway is that seemingly absurd growth rates by historical (empirical) standards are rational expectations that have been born out by our SaaS stocks. The corollary is that the fast growth stage is compressed in time because it’s market size that limits growth.
To conclude, the table I posted covers 25 scenarios depending on
-
Share of the market the company in consideration has of the technology it’s engaged in, and
-
Effective dollar retention rate achieved by the company.
The table was an effort to show the possible futures of ZScaler not an 85% prediction of anything.
The Captain
48 Likes
Having lived in the valley all my professional career, it is difficult to not know ‘S’ Curve :). Couple of points to consider
- Not all technology gets to ‘S’ Curve;
- The timeline varies; or aided by external factors (like other complimentary technologies adoption/ growth)
Now, you may think the growth of SaaS applications requires or even mandates a complete paradigm shift. I view the bulk of the applications I would say still 90% reside within the DC or customers own “firewall”. Now, this migration could accelerate in the coming years, and necessitate parallel, complimentary approach, i.e., you need firewall appliances at the edge of your network/ DC, and one out their in the “cloud” to handle all “internet bound” traffic. Of course you could assume ZS technology or the “architecture” is so superior that it will completely eliminate appliances.
Where we are today, I see appliances are going to be there for foreseeable future. The adoption of “cloud” and the applications moving to “cloud” is critical for the migration from “appliances”.
In a nutshell, migration of applications to “cloud” en mass is going to be critical driver for the technology to get to 85%. I understand you believe “cloud” makes things “easy”. I live in a world where whether it is “social security” or “medicare” or “Aerospace” or “defense” vendors developing super-high technology products are still running their critical applications in “COBOL/ C” or whatever legacy application with which they started.
To put it in your language, you are going to have a long period of “J”, and the evidence of it will be growth slowing to 30’s, 40’s. Your short period or 5 year period “J” would require the industry to grow above 75%.
I don’t see that now. Hence I am doubtful about the 5 year period.
Don’t get me wrong a decade long 30’s to 40’s growth is not a bad thing. It is in fact a terrific business to be invested in, but that is not same as getting to 85% adoption in 5 years.
3 Likes
I live in a world where whether it is “social security” or “medicare” or “Aerospace” or “defense” vendors developing super-high technology products are still running their critical applications in “COBOL/ C” or whatever legacy application with which they started.
It is absolutely true that critical legacy applications, based on old tools, are still critical, and aren’t going to be replaced any time soon. However, critical is not limited to legacy. As information becomes continually more important, many of the newer applications - based on current technology - become increasingly critical.
5 Likes
It is also true that these legacy critical applications have remained in their ancient technology in part because replacing them with applications in new technology seems overwhelming. But, development productivity continues to improve … at least in the leading edge … and requirements change … making a COBOL application accessible from the web can easily be more difficult that redeveloping the whole application in modern tools that provide web access. So, while the persistence of these legacy applications is pretty astounding, that could change fairly quickly without any new technology needing to be developed. The killer, of course, would be a tool that would translate the COBOL code to a modern language, but so far this seems elusive and may remain so because COBOL applications, even optimally structured ones, are structured very differently than is optimum for modern languages. If it takes one as much work to restructure the application after translation as it would have taken to just write the application in the new technology, then one hasn’t really gained anything by the translation tool and the result is actually likely to be less ideal. Model-Based Development has promise to make such transitions more efficient, but seems to be slow to catch on.
1 Like
I view the bulk of the applications I would say still 90% reside within the DC or customers own “firewall”. Now, this migration could accelerate in the coming years, and necessitate parallel, complimentary approach, i.e., you need firewall appliances at the edge of your network/ DC, and one out their in the “cloud” to handle all “internet bound” traffic.
A few points.
-
If you have any cloud applications than you need a solution to secure that application. So yes, I agree the old paradigm of a firewall was the approach to old applications. The reality is most places are phasing in cloud applications so even though many applications can live behind a firewall, most users need the ability to access cloud applications. If you are going to go through the trouble of setting up zscaler why keep your appliance?
-
The idea of a perimeter securing your network is long gone. Zero trust is the name of the game. I don’t know enough to say that zscaler’s approach is definitely better but the idea of application/user-specific authentication (zscaler) makes more sense than trying to bolt on zero trust to an applicance(PANW).
-
Remote access is becoming more and more important.
-
Forklift upgrades vs known monthly fees. We are seeing monthly fees winning out because it makes budgeting easy.
best,
e
19 Likes
If you are going to go through the trouble of setting up zscaler why keep your appliance?
Clearly right? Now, flip the question other way around, if 90% of my applications are going to be within the firewall and the appliance is taking care of it, for 10% of applications am I going to completely rip and replace everything?
The way I see it, until a sizable part of an organizations applications move to cloud, they are going to be careful/ reluctant to rip and replace their existing “architecture”.
The way I see is, the current adoption levels are not telling me that the appliances are “ready” to be phased out.
I guess that difference of perception is what makes the market.
3 Likes
ZPA removes the need for any DC appliances in regard to serving out internal apps.
Depends on the end as to pain point. In the Zscaler presentation a multinational oil conglomerate gave a very compelling presentation on multiple points in the context of oil prices falling from $100 to $30 a barrel.
(1) appliances are sold with a 5 year life cycle and annual maintenance contracts. The company could not afford redundancies and therefore if an appliance went down it was by passed and there was no security.
With this 5 year life cycle it gave the company the impetus to say “no more, not one step further, get rid of the appliances!”
(2) the company could not afford bolt on SSL inspectors. As Zscaler has stated (and Palo Alto omitted) appliances cannot read the bulk of SSL.
Turns out the bulk is security threats were SSL when they finally were able to inspect them. Their monthly threat counts were down something like 25x thereafter to almost nothing.
(3) when they acquire a company suddenly bandwidth demand goes up. The 100gb boxes they bought are not enough. They need to upgrade to 250gb boxes. So what to do with the 100 go boxes?
(4) an executive cane complaining that he had better download speeds in 1995 with dial up. And the tests proved that he did than with the appliances as security.
And I could go on. There has not been a single patch that he is aware of (as Zscaler takes care of it) and bandwidth flexibility is invisible with Zscaler.
An example from another customer. They did not even know it but they had a 3x bandwidth explosion when the CEO gave a utube address to the company. No one even noticed the spike. It was such a remarkable occurrence they made a business study out of it.
Palo Alto’s Prisma Access cloud product requires that each virtual firewall (as that is what the product is) be manually set for increased bandwidth. And I the process multiple variables such as IP addresses may need to be reset for each virtual firewall.
I mean, seriously there are many more examples that appear over the top but are true.
I have no doubt that many will go screaming into the night before giving up their appliances. It is going to happen though. When? We shall see. Which company replaces them? Not a difficult question at the moment. Will more real competition rise? I don’t know. What Palo Alto is doing is creating a cloud offering to enable them to maintain and sell more appliances. Their offering cannot stand alone and compete. It only works with an appliance infrastructure.
For whatever it is worth.
Tinker
27 Likes
Yest Salesfore only has 20% market share in CRM. Still Siebel is used by all major US airlines companies for loyalty program, same for major consumer product companies, etc and the install base of Siebel more or less survived, CRM took lion share of the new implementation but was not able to dislodge the install base.
I think the install base survived with companies that mostly use Oracle as the ERP. Most instances of SalesForce I have seen are integrated to a back end ERP like SAP or Oracle. Those usually are source of things like the customer sold to’s, material masters and price books that feed into SFDC. I have seen companies that are using Oracle have Siebel running as part of the ERP that maintains a lot of the backend function to feed it to SFDC which then is used by sales. In those cases, Sales has no idea that Siebel is actually in the background as part of the ERP to support SFDC.
I am sure a large part of Siebel’s install base remains to help support SFDC which is being used by the Sales team. In those cases, they did dislodge Siebel as the user facing CRM but it maybe performing background functions that Oracle doesn’t do as well. In cases of SAP, usually I see SFDC tied directly to SAP without Siebel
1 Like
I mean, seriously there are many more examples that appear over the top but are true
Oil wells now need to communicate back to HQ, so the oil companies are one of the early adopters. Lot of interesting issues. But, you know they are not compelling reasons to believe “appliances” are going out thesis. They may get replaced eventually but not yet. Like I said lot of difference between 30~!40% grower vs 85% market share, I am speaking about the technology.
1 Like
Kingran
Like I said lot of difference between 30~!40% grower vs 85% market share, I am speaking about the technology.
Without wanting to speak for Denny but I believe he was talking about 85% market penetration of the technology adoption as opposed to 85% market share for ZS.
A
3 Likes