Gerardo Dada

01:25 - 01:26

Alright. Well, we'll.

Gerardo Dada

01:26 - 01:56

get started now. Well, thank you everybody for joining us today.

Good afternoon for a good morning if you're in another part of the world. Today, we're gonna talk about why IT must rethink monitoring for the Cloudera.

And, specifically, we're gonna talk about the the maturity model with GigaOm. We are honored to have Howard Holton, COO for GigaOm, and, one of the principal another principal analyst for GigaOm.

Thank you, Howard, for joining us today.

Howard Holton

01:56 - 01:58

Thanks for having me.

Gerardo Dada

01:58 - 02:07

And we also have Mehdi Daoudi, industry veteran in monitoring and also CEO and founder of Catchpoint Systems. Thank you.

Mehdi?

Gerardo Dada

02:07 - 02:12

Cofound cofounder, so that way I don't make enemies with my other cofounder, Gerardo.

Gerardo Dada

02:12 - 02:13

That's great.

Gerardo Dada

02:13 - 02:16

Thank you thank you so much. Welcome, everyone.

Gerardo Dada

02:16 - 03:15

Yeah. And this is Gerardo.

I'm I'm the field CTO for for Catchpoint, and, very happy to be here with with this is gonna be a very interesting webinar with two people who have a lot of experience to share with you. We want to make this session very interactive.

So if you have any questions, there's a q and a tab. You can also download a copy of the GigaOm maturity model in the in the top tab at the right side of your screen.

And with that, we'll get started. The the topic today is how monitoring needs to evolve and why do we need and how can companies today use a maturity model.

So first question I would like to ask is, what has changed in the world that requires organizations to change how they think about monitoring in general? Right? Like, monitoring has been on for forever since since IT began. And traditional tools like APN, they're already 30 years old.

Every company uses them already. They've matured tremendously.

Why do they need to update the tools, the processes, the way they think about monitoring? Who wants to go first?

Gerardo Dada

03:15 - 05:37

I'll go I'll take a stab at this. So I think three main things have changed, And I've been doing monitoring for almost thirty years now.

I'm aging myself, but I think my haircut would would would imply about it. The first one is the requirements for speed.

I think that is or the requirements for performance. The requirements for resilience has changed.

In the nineties, you were okay downloading a web page in ten, twenty, thirty seconds. I think today, your web page, your web application, your mobile app needs to run at speed of light.

So I think the the the the requirement from the user perspective have changed tremendously. The room for error is, unforgettable.

You literally would would lose customers and and and and everything by not having a system that doesn't work. So that's number one.

The second is the complexity has changed. I think we went from single data centers to multi data centers to multi cloud, multi CDN, multi disk, multi DAT.

And so from a complexity perspective, everything is exploding, and I don't see that going simpler. Actually, it's getting more and more complex.

Now you're going to add chat chat GPT bots and MCP servers to add more complexity. So I think we are in an exponential race for more complexity.

So complexity are still requirements for performance and resilience. And then the the the third thing in my opinion is just the speed at which we we must process all this data that is coming from all these monitoring tools and systems and just the signals that we're getting is getting is is so much that the human brain cannot comprehend all these things at the same time and then connect the dots.

I think those are the three things where I or we almost need the not a reset, but, like, a step back. It's like, okay.

What are we doing here? How do we how can we deliver on our business requirements? Because, again, remember, we're all need doing these things because at the end of the day, we need to help businesses run better. We don't do monitoring for the sake of monitoring.

I don't think that, anybody will will fund those those projects. So, those are the, you know, the three things that I think have changed tremendously over the last two, three decades.

Gerardo Dada

05:37 - 05:42

Thank you. Howard, what's what's your perspective? Anything you want to add or any different perspective?

Howard Holton

05:42 - 07:53

Yeah. I mean so so I I think to just kinda double click a little bit.

Right? You have two hundred milliseconds. That's what you have.

You have two hundred milliseconds to show activity back to the user so the user recognizes that your website is doing something. You don't have to have full page loads, but you have two hundred milliseconds to show that your application is performing in a way that people feel that the application is live and working correctly.

That's a fifth of a second. That is not a lot of time.

Right? To to Mehdi Daoudi point, this isn't 2,008 where you could get away with a two or three or four second load time. Right? So you you really have to respond very, very quickly.

Additionally, the distance between the things you can control and your users has increased exponentially. Things like work from home really changed how we view, the boundary of our network and ultimately added this enormous amount of elasticity between us and our our own corporate users through a series of networks that we no longer control, which makes it really critical that you're able to find out where any issue lies and be able to quantify it in a way that's that's measurable and reportable back to your leadership.

And it can't just be, well, I don't know why that is now seven seconds. You really have to be specific.

Also, the telecoms have recognized this and provide methodologies for you to report. Whereas, you know, when I when I started tracking this stuff back in 02/2003 and 02/2004 working for for WellPoint, you kinda just went, it doesn't work.

Right? Like like, I did a freeze route, and it's broken it in some path, but but now we're lost. No.

Today, you can actually get to the telco. You can report the path.

You can you can report it, and they will take action. Right? We have amazing route routing protocols that will help to that will help to solve some of these things, but you really have to stay on top of it.

And it is it is kinda to Mehdi's point. Right? It is entirely too much for any person to keep in their head, which means you really have to tool up.

Gerardo Dada

07:53 - 07:53

Yep. Yeah.

Gerardo Dada

07:53 - 08:40

You know, one sorry, Howard. One thing, Howard, you said that is, is so important is there is no more, excuse in terms of saying, I don't care.

It's not my problem anymore. I I remember in in we we used to be able to go back to our CEO when I was at double click in ninety seven.

It's like, well, you know, it's the Internet. Go figure it out.

Well, today, that that doesn't lie. People you are you're expected to find a solution.

Raft around it. Find another provider, etcetera, etcetera.

So the days of going and hiding under under a rock and hoping that the problem go away, The stakes are too big, so there is no no more excuses. Sorry.

No.

Gerardo Dada

08:40 - 09:09

No. I was gonna actually make this similar comment where where it's not about the websites anymore because now every business is dependent on the Internet.

Right? So twenty, thirty years ago, you worked for a web down, like, man, nothing happens. Nowadays, the connectivity goes down, your your business stops because everything is dependent on Internet and cloud technology.

So we we mentioned website response time, but it is not only about websites. Right? There's about applications and APIs and involved in interconnectivity.

Hospitals. Hospitals, factories,.

Gerardo Dada

09:09 - 09:31

food delivery, you know, food manufacturers, every millisecond count. And and, and I think it's all again, it's just like how digital the world is.

I mean, it's funny we still talk about digital. I don't think of anybody's going, hey.

I have an analog, company. Can can you please help me? Right? That doesn't exist.

So everything is 10001000% digital.

Gerardo Dada

09:31 - 09:33

Well, it's it's not it's.

Howard Holton

09:33 - 09:35

never gonna go the other way. Right? Like.

Gerardo Dada

09:35 - 09:35

Correct.

Howard Holton

09:35 - 10:29

The the the truth of the matter is while we have talked about the application layer and the thing that affects users, what what what you hinted at, what you kinda touched on, I guess you didn't hint, you touched on, is there's this whole sublayer of multiple pieces of data stream data fabric that's constantly streaming in that provides all of the context around everything that our users are doing that's also really critical. Right? Things things that we use for cart abandonment, things that we use for user abandonment rates, things that we use for user activity, Those aren't things that affect the user, but they're really critical that we're able to receive them receive them in a timely fashion, and that those things are correlated to what's actually happening.

And when you have four or five different data streams going on, if those data streams aren't correlated, those data streams aren't received in timely fashion, process in a timely fashion fashion and action taken, that opportunity is lost and that opportunity gets lost forever.

Gerardo Dada

10:29 - 10:44

Yep. So so in the face of all these changes and how technology has changed, how how should organizations think about evolving their monitoring strategies? And what are gonna be the most common pitfalls and challenges they have in that process?

Howard Holton

10:44 - 11:44

So I would say the first and and foremost thing is you need to start thinking about your monitoring, and how does your monitoring represent your user experience. Right? If you look at traditional monitoring, traditional monitoring did a fantastic job when we had complete control over the boundaries of our network.

When everyone was on prem, right, we had complete control over the network. And, really, if we're being honest, we didn't spend a ton ton of time thinking about the network.

It just kinda sorta worked. Right? I mean, how long did we stay with endpoints at one gig? Mhmm.

Right? And and and we've moved past past one gig from a technology standpoint a long time ago. Right? Most of the companies I talked to, their data centers were still 10 gig for a very long time.

Right? And so you concentrated mostly on the application stack and the things that made up the application stack. And our applications were pretty, you know, were pretty pretty, siloed and pretty, what's the term I'm looking for? Help me out here.

Gerardo Dada

11:44 - 11:45

They're self contained.

Howard Holton

11:45 - 12:00

Yeah. They were self contained.

Monolithic. Thank you.

They were they were they were monolithic. Now the application isn't monolithic.

Now all of the application traffic happens over the network. Right? So so then you make that again percent.

Gerardo Dada

12:00 - 12:07

And, usually, 80% of it outside of your own bound Right. Even your your your fake boundaries.

Right? I mean.

Gerardo Dada

12:07 - 12:09

Right. And you look at the web APIs.

Howard Holton

12:09 - 12:55

that we're using. Right? You look at the APIs that we're using that that allows us to support the application.

How many of the applications can we actually say 100% of the application value exists within our network? It's not overly high, and and those certainly aren't the modern applications. So you have to look at monitoring the same way.

How do I make sure that I'm monitoring in a way that that allows me to closely replicate what my user's doing, what my user's seeing? Well, I can't do that on the networks that I control. I have to leverage monitoring solutions that are cloud based in their nature, but also distributed in a way that allow me to really get a clear picture of what's happening, what's happening in the world, where is it happening, how is it happening, and how does that match our our users and our user profiles to create that really clear picture of what reality looks like.

Gerardo Dada

12:55 - 14:54

So, Howard Holton, I I have this slide that I've used for the past fifteen, sixteen years. So when I was running the monitor the the the monitoring team at DoubleClick, I had a budget of about $1,015,000,000 dollars a year.

And I had the pyramid that started with infrastructure monitoring. That's where most of my money was spent.

Then it would go network. Then I would spend on on QA testing.

And then a tiny, teeny, tiny little bit was spent on Gomez and Kino to the grandfathers and grandmothers of end user monitoring. And so where today, that pyramid, that triangle is completely reversed.

You need to spend I would have spent most of my money on all the things that we look at from our side perspective, monitoring your CDN, your DNS, all these critical third party services, and then go all the way down to the stack where infrastructure monitoring is, almost like a commodity to some degree because that's provided by some of the cloud providers and whatnot. But that is less relevant because, again, the users are less concentrated into that monolithic thing.

That thank you, Gerardo. He he, by the way, designed that on the fly the last two seconds.

Right? So this is the issue of Gerardo. It's better than said GPT, but, this is exactly the way we used to do things, before and after.

Right? So it's like, how do you bring a an end user mindset to monitoring? And I think that's the biggest transformation that we've been preaching for the past sixteen years, and that I've preached even internally at DoubleClick is, like, you know, my network was always fine, but our we were slowing, ad serving, or we had a lot of problems, serving ads fast because we were not looking at it from an end user perspective. Well, our data center is fine in in in in in in New York, but then, you know, what happens with the people in Kentucky? How fast were they getting the ads? And so we have to shift things and look at it always from an end user perspective.

‍

Gerardo Dada

14:54 - 14:59

Great. And so in in the context a really good point.

Sorry. No.

Go ahead.

Howard Holton

14:59 - 15:08

That's a really good point. Right? When you look at at kinda how this stuff was all built, it goes back to the days when the infrastructure just wasn't reliable.

Gerardo Dada

15:08 - 15:09

Yes.

Howard Holton

15:09 - 15:44

Right? We spent a lot of IT time making sure that things didn't overheat, making sure that hardware didn't fail, making sure that that the hard drive spinning rust stayed spinning. Right? We we were more concerned about blue screens of death and application exceptions than we were quality of service.

Right? Because because, ultimately, the stuff wasn't reliable. Today, like, I have I have stuff running in my home lab that if I built a wall, I could leave it running for twenty years.

The hardware, frankly, today is super reliable, much less I mean, that's the stuff that runs in my home lab. Think about the stuff that runs enterprise wide.

Right?

Gerardo Dada

15:44 - 15:44

Yeah.

Howard Holton

15:44 - 16:02

So I'm not so concerned with, is the infrastructure capable of a 24 by seven three sixty five duty cycle? I'm not so worried about that. It's much more service availability based on the thousand other variables that occur.

Right? And and this is the most important complex.

Gerardo Dada

16:02 - 16:37

And you're weakest link, Howard. Right? So you might you might also design and build the greatest system in the world with the greatest hardware in the world, blah blah blah.

But at the end of the day, you are as strong as your weakest link. Right? I mean, we see that in security.

We see that in infrastructure. We see that in resilience.

We see that at the airports. You name it.

Right? So this false sense of security that, oh, I'm I'm the strongest. I'm the best.

Actually, no. Did you did you leave a little door open of something that can literally wipe you out in two and a half seconds?

Howard Holton

16:37 - 16:39

Yeah. %.

Gerardo Dada

16:39 - 17:17

Great. So as as all these things change, like, if we're no longer, you know, monitoring IT operations, it's no longer about, CPU utilization, right, or server health, which used to be kind of the primary concerns of monitoring systems even, I don't know, fifteen years ago.

Now things are changing to monitoring those dependencies and those APIs. That that level of complexity that we, we we call the the Internet stack.

Right? So how does this change monitoring, and how does, you know, what does this mean for for IT operations? What is the how do you monitor this Internet stack?

Gerardo Dada

17:17 - 20:02

I think the first thing is and I I don't want to come as too biased. First is because of these layers is the first thing is the biggest mistake I see people make is believing and thinking that one tool can do it all.

Mhmm. And that, unfortunately, is a a again, I've been doing monitoring since the nineties, and I've been promised by every vendor out there that there is a single pane of glass that I can go and buy and magically is going to tell me everything I need to know in one shot.

And, unfortunately, there are some people that believe in that, and, you know, they are on a journey. Hopefully, it won't be too the the landing won't be too hard, right, when they realize that that doesn't happen.

But when you realize the complexity, it's the same thing as a human body. Right? When you look at our us as as humans, we have a brain.

We have skin. We have we have, miles and miles of of veins.

We have heart. We have mechanical stuff.

We have software. There isn't a single doctor.

I don't know. I I mean, I was joking with someone the other day.

I don't know when was the last time my dentist gave me a colonoscopy. But they are both doctors.

Right? They all went to medical school. Right? They're all a monitoring tool, but I I don't know.

And maybe the drilling, but so the the biggest mistake can people make is by understanding that I'm going to just there is a magical tool that is going to do that. Then the second thing is, again, I think people need to take a step back and really understand in that stack what is going to hurt them the most.

And I think then deploy take a step back, then deploy the right monitoring strategy. Again, strategy, not tool.

The right monitoring strategy to understand your risk Because monitoring is about risk. Mhmm.

I'm going to use a technical term here. Shit is going to happen.

What so things break. This is the reality.

And monitoring is about ensuring that resilience. You see behind, on top of my head, there is a smoke detector.

I'm in a hotel. They could have just you know what? While I spend money, put a smoke detector in every room, maybe we just put one on the First Floor.

No. There is one in every room because you want to catch something before it's too late.

Right? Because the sooner you catch something, the easier, the cheapest, etcetera, etcetera, is it to fix. So I think there is not a single tool, understand what can hurt you, and then deploy a monitoring strategy that is going to allow you to catch a problem as quickly as possible.

Then go and buy the tool or tools to be able to alleviate that particular risk. But monitoring for me is risk management.

That's all this.

Gerardo Dada

20:02 - 20:04

is. Yeah.

I agree. Right?

Howard Holton

20:04 - 20:43

That from an IT operation standpoint, it what the business wants is for IT to tell the business when something is down, not the business notify IT when something isn't working. Right? They really don't when you're looking at organizations whose common theme is what do I pay you for, well, that is specifically what they pay us for.

And when they're constantly coming to IT going, why isn't this working? Rather than IT going, hey. We're we've detected an issue.

We're aware of the issue, and we're working on the issue. That's all they want.

Right? They just wanna make sure that someone is paying attention to it whose job it is to pay attention to it.

Gerardo Dada

20:43 - 21:52

Howard, and, you know, for me, the best companies is when people have already understood that, but then here is where the tech is when IT becomes a business partner and they say, I can run the business. I can help you run the business better.

I can improve ourselves in China. I can improve ourselves in The Middle East.

I can I can hey? We need I think we're getting enough traffic from Saudi Arabia. It's time for us to go and maybe evaluate an infrastructure in Saudi Arabia.

Or, hey. The reason why things are not working in Saudi Arabia is because our CDN doesn't know how to serve Saudi Arabia correctly.

I'm using that because it's a real example. I just dealt with.

So so the yes. The first one is you want IT to be preemptive and know about the problem before the business is like, hey.

Why why why are we down in Germany? Oh, you didn't know. Why do I pay you for? Then the second thing is when the maturity is when actually, for me, at the top of the maturity is when IT becomes a key player.

Somebody at the business table saying this is how we can run the our businesses better. We can save money.

We can do more. We can do more with less.

That's when things change in my opinion.

Howard Holton

21:52 - 22:33

Oh, I I I could not possibly agree more. Right? The truth is we exist to serve the business, and the best way to serve the business is help the business achieve its goals.

Its goals are not buying technology. No business is in the business of buying technology.

They leverage technology to create value. So how do we help the business create more value? We help by getting in front and by providing insight from all of the things that we do.

So if if step one is make sure that you're proactive, not reactive, then step two is make sure that you're providing insight and helping to drive the business forward. If you don't feel that your IT department is welcome at the table with the business, These are probably the reason why.

Gerardo Dada

22:33 - 23:23

Correct. Correct.

There is this, cartoon I've used in the past, which is like a the CIO organization is angry because they're not felt they're they're they're not feeling included at the table. Right? Well, it's because you are not helping business run better.

Mhmm. And so I have so many customers that when they achieve that, when they actually turn the table around and they can run their businesses better I mean, I have a customer that, was having a lot of downtime over the years, and the leader went and took a complete different approach where the last two years, they've had zero downtime during their ecommerce period.

And the the the person is being promoted because they are running the business better. Right? And that's what you want.

It's like, I am sitting at the table because I'm adding value. I'm not just a cost center.

Gerardo Dada

23:23 - 23:56

So that's at the heart of the maturity model, right, going from from reactive to proactive to predictive, going from focusing on the hardware to look seeing systems to work for cleaner business. You get customers like IKEA where their monitoring team is now called operational intelligence team, right, because they see themselves less as an IT function, more as a system of how the the business works, but just everything is IT related.

So how how is this related to the the GigaOm IT maturity model, and how should organizations think about it as as a framework?

Howard Holton

23:56 - 24:30

Yeah. So when we looked at the maturity model, right, to make it simple, we really separated into kind of four focus points.

The first is technology focused. Right? So so everyone kinda starts at a technology focus, and you kinda have to.

Right? Here, you're looking at stepwise improvements, system level indicators, some technical KPI tracking, and, really, you're focused on detect and restore. Then we move to system focused.

Right? And so we change from kinda siloed operations to integrated operations. We're now we're really looking at being preemptive.

We're focused at system level objection sorry, objectives.

Gerardo Dada

24:30 - 24:31

Objectives.

Howard Holton

24:31 - 26:55

We're focused on result tracking, objective measurements, and a mean time to identify. Right? These are much more system focused.

Now we move to experience focused. And this is where observability changes from kind of a minimum viable observability, right, which really isn't focused on the business, but is required for technical teams to be able to do their jobs and react properly.

Now we're focused on value based observability. So now we're really looking at what is the experience.

And the experience, all of a sudden, becomes open and collaborative from a culture standpoint. You've unified the operations mission.

You've got high resiliency in all the domains. It's an IPM driven design, and your decisions are data driven.

They're data driven actions, and they're data driven improvements, which finally leads us into kind of a value focused observability strategy, which is continuous improvement, innovation led. And now what we're doing is we're looking at optimizing digital experience across the board, really improving brand protection, enhancing workforce productivity, and prioritizing ultimate resilience beyond just simply monitoring.

This is the point where you start to become quantitative and and optimized within your maturity, and you're really driving value back into the business. And the amazing part is all of that data comes together to do things like Mehdi was talking about, where you can see a really granular look at what's happening within every country that you serve and start to take that back to the operations team and go, hey.

I've noticed that within the last six months, we've seen a big surge in customers in Saudi Arabia. Are we doing anything specific with Saudi Arabia? Oh, no.

We should probably think about it. It's our fastest growing reader based on Internet traffic.

Maybe we need to advertise to Saudi Arabia or being able to go, hey. We've seen a 45% increase in European operations.

How is our GDPR? Do we need to be concerned about our use of AI? Are we following those compliance regulations? Because it's entirely reasonable that the CLO is not aware, right, that something has changed. And and all of a sudden, IT becomes the nexus for this this communication within the organization, and all of that is value driven.

All of that is providing value back to the business. And then all of a sudden, the question goes from how can we save money in IT to holy crap, this is hugely valuable.

Now it's a place where I can invest and get and get an ROI. How do I invest more to get more of this kind of thing?

Gerardo Dada

26:55 - 28:14

You're you're so spot on. When, when I was running the quality of service department at DoubleClick, the data became so critical in running the business.

I was never asked to cut our budgets. We were always asked, what else can you do if I give you more? What else what other insight can you give us when and that's what I see with some of the customers we work with is when they turn that I mean, we we've known this.

We we talk about data is the is gold. Right? Data is the new oil.

Right? Well, yeah, data is the new oil because it can run businesses better. But we need to move from being reactive to proactive.

Right? That we need to put the fires. We need to start thinking about frameworks where we can use all that data.

And it's about finding those gold nuggets. Right? And that's ultimately what observability is or should be, right, or monitoring or whatever you want to call it, is how can we run our businesses better.

And I you know, it's, but it starts with, like, what it starts by saying we are here to help the business. I think the biggest transformations I've seen with SRE organizations and IT teams we've I've been work personally working with is when you have a leader, people that are saying, I'm going to be a business person first.

I'm not just the bits and bytes. I'm actually here to help.

I'm I'm the oil that makes a machine run better.

Gerardo Dada

28:14 - 28:33

And in in this context that you mentioned cost savings and budgets, maybe maybe a question for Howard. How can IT leaders get the buy in from organizations to make the investment required to modernize that solubility and advance your maturity model in order to start talking about saving money and cost cutting costs and reducing headcount?

Howard Holton

28:33 - 31:04

So talk about the value to marketing. Talk about the value to sales.

Talk about the value to operations. Right? We talk about a single pane of glass.

I absolutely hate single panes of glass. A single pane of glass says everybody does the same job and everybody has the same need, and that is never true.

Right? But if you have a good monitoring solution that looks at the globe, you have an a number of reports that are hugely valuable. Forget about dashboards for a minute.

Not everybody needs to know what's happening minute by minute. But month to month, there's a tremendous amount of value in being able to validate what operations thinks is happening, validate what marketing thinks is happening, validate what sales thinks is happening, and reaffirm their own data.

Right? So all of a sudden, this this this isn't about IT having having, another tool in its toolbox. This is about the entire business leveraging data as a check to what they already think is happening.

We know this is happening because we can see and observe the data, and now this becomes a feedback loop for marketing. Hey.

Are those marketing efforts making sense? Are we putting the money in the right place? Are we seeing that return? Right? Sales. Do we have the right sales motions in the right place facing the right places? Even little things, like being able to go, hey.

We've seen a a huge uptick in China. First, do we wanna see the uptick in China? Second, how are we serving the Chinese market? Is it time to start talking about some translation and some localization that maybe we're not actually doing? Right? Maybe we look at, are we are we are we as a culture, as a company culture, able to serve a market that maybe we don't understand so well? These start to become large scale strategic value points that the data is so clear and so easy to get to within a good monitoring solution that this stuff starts to become really, really clear.

And then, oh, by the way, this also helps us reduce risk. Oh, by the way, this also helps us protect the brand.

Right? But there's so much value here that I I really think you have the ability to pinpoint exactly where the organization wants to go. And, frankly, the best way to start is take a look at your company's 10 k.

Take a look at your company's shareholder and and financial analysts, briefings. A lot of those are public.

Right? Listen to them and see what are the things that that they talk about as missions and goals for the company, and then just align your your messaging around it and keep it really simple. You you can I promise you can do it in one slide, and you can find the trigger that's going to work that's gonna that's gonna help move the needle forward in this space?

Gerardo Dada

31:04 - 33:29

I'm laughing because I had this, contentious meeting a few years ago with a customer who was literally shutting down all their observability, all their monitoring platforms. And I and and while I was walking to the conference room, there was a picture of the CEO that was talking about this year, we're go going to focus on customer experience.

I said, hold on. Explain this to me.

Your CEO is saying everywhere that we need to focus on you guys need to focus on customer experience, and then you're telling me you don't care anymore about performance? It's just like it's like and and the person didn't have anything to say. What I will add here is, like, for those that remember their marketing one zero one classes in college, so marketing, the the four p's of marketing, price, promotion, product, and placement.

Right? These are the four p's of of of of marketing. How you price, where you promote, how you promote, where you place, and and the product itself.

But there is the I've been saying this for a long time. There is the fifth p.

There there there is a fifth p in in in in the new sheriff in town. And the world we live in, the p is the performance.

Right? So the reliability, resilience, you can call it, what does it help with? Well, it helps with performance. Right? And so in today's world, you cannot neglect that fifth piece.

So, ultimately, you know, when you're negotiating when I was negotiating my monitoring budget, I mean, one of the biggest downturn I was part of is February, the. com boom.

Right? And and one thing we never cut, and I would I would never forget that our CEO was at the time was was brilliant. He said, we're not going to cut on monitoring because all the other companies, all of our competitors are going to cut on monitoring.

Their quality of service is going to tank, and we are going to double down on performance because we're going to be around when the mark because everything is cyclical. Right? When things come back up, everybody's going to know that we were the best, the fastest, the most, most resilient ad serving ad tech company.

And that's exactly what happened. Right? And so especially when things hit the fan, I think this is where people need to double down because your customers expect you to.

They don't expect you to you know, they don't expect, like, Apple to shut down their website and say, you know what? Yeah. We don't know the tariffs, this, that, whatever.

How about we just close shop and wait or Costco or any of the companies? No. This is where people need to to double down, on their efforts to remain resilient for their customers.

Gerardo Dada

33:29 - 33:45

That's great. And so maybe the next question is, so we have this maturity model.

How should companies use this maturity model to guide them in this progression? What is how how do you recommend that they use the maturity model, operationally? How do they implement it?

Gerardo Dada

33:45 - 33:51

So the way we've been think yep. Go ahead.

Who was it for Howard? Yeah.

Howard Holton

33:51 - 35:22

So so my advice with with the maturity model is is keep it simple. So maturity model is designed to have two points on it.

The first is figure out where you are today, right now. It's not about making a judgment call.

You have to know where you are, and you have to be really honest with yourself on where you are, what your capabilities are, and where you kinda slot into the model. And then look forward eighteen months and see where do you wanna be.

Right? So if you're a maturity kinda kinda sitting at one, one and a half, almost two, and you're like, but we'd really like to be in that value based space. We'd really like to be providing that back to the business.

We'd really like the business to get the value of these things. Well, then you need to move to a three and a half or a four, maybe move to a four to a four and a half.

Then it becomes really easy to go, okay. Cool.

What do I need to do get there? How much of that can I do internally? And how much of that do I need to build an ecosystem of partners to help me get there? This stuff isn't easy, and it's not it's not something that you can generally DIY. And that now now that gives you the framework for a strategy.

Now you can start building a strategy. Now you can start talking to an ecosystem.

Now you can reach out to companies like Catchpoint. Right? Companies that are experts in this and go, hey.

This is where we are. This is where we need to go.

You've now created a wonderful framework for an entire conversation. Help me build the strategy.

Help me get there. And now you're having a really good conversation that has some some clear goals, some clear boundaries, and a clear outcome already decided, and that's what a maturity model is really designed to do.

Gerardo Dada

35:22 - 35:39

Maybe maybe I totally agree with the Yeah. I'm sorry.

What what is the role of automation and AI that it plays in advancing IP and mature? Or, like, AI is a is a almost a buzzword nowadays, but automation is also important in the in the context of being efficient and being more mature.

Gerardo Dada

35:39 - 39:41

Yeah. But I also think that it depends where you're on the maturity model.

I think if somebody's in the crawl space, right, at the crawl phase, I think throwing AI at this stuff might sync them. Right? So, like Howard said, this is your journey.

I think people need to first agree that they need to get on the on the bus. They need to go from point a to point b.

They need to be they need to be honest with themselves where they are. I mean, we have some customers that are at the beginning of the maturity, and all they need to improve is their availability, their uptime.

Great. Well, start there.

But I think at the end of the day, as I said, there's a lot of data that we collect, that the other tools are going to collect, etcetera, etcetera. I think AI is going to drive two main things.

One is automation, being able to do real time things, but that's going to be reserved at the beginning for very highly sophisticated companies that are having multi CDNs, multi cloud, where I I have, like, for example, I might be using Akamai and CloudFlare. It's Fastly as a CDN.

I want to be able to do real time, translation of where to send my users in China or my users in Pakistan if there is a problem. I want to be able to send, hey.

This shopping experience needs to go through through an edge work on CloudFlare or something else on Fastly or maybe on GCP. I think that's where AI is going to help is drive that level of automation that is going just take the human element out of it once you do that.

I think that's where I see the AI power for for for us. The other thing that I see is, like, once you are maturing, you're going to develop better playbooks, better runbooks, better understanding of your infrastructure.

And I think where AI is going to help is is help you run those playbooks and make sure you didn't forget anything. Right? Did that did I miss a spot? The third area where I just think AI is is just so much better, at this stuff is connecting the dots.

Because if if if you agree with me that there is not a single tool that has the ultimate knowledge of everything that is running, but you're going to have you have to you have to come to to the realization that in order to for you to run a better business, you are going to collect data from various probes, various systems, various monitoring tools, and the business tools, and and just not IT. But you should you should be able to gather a ton of telemetry from from a lot of different, providers.

And something is going to have to connect the dots in high speed at high speed, and that's where I think AI is going to shine in in our world. Right? It's it's being able to look at your logs, your traces, your synthetic data, your run data, your Glassdoor data, your your your business, your your your your payment system metrics, all that stuff.

And then you say, hey. We have a problem in Germany.

Why is that? Or we have a problem somewhere else in Brazil. Oh, the Brazilian government just forbid x, for example.

Well, that had an impact maybe on revenue. But you cannot do that just with the IT telemetry, without the business telemetry, etcetera.

So so, I mean, some of the some of the most advanced customers we work with that are on the high end of the maturity are not looking at the data in Catchpoint or any of the other tools. They're pouring their data into into the Databricks of the world, into the the the snowflakes of the world where now they are able to literally correlate and look at the root cause analysis and just connect the dots across multiple multiple things.

You know? I'll end with this. We went from looking for one needle in a haystack.

So this imagine this. Now you're looking for a needle in multiple haystacks.

Right? And and this is the the challenge that we all have to deal. And this is not just an IT guy or an SRE guy or whatever.

The business too need to run better, and they have to answer those those questions. So so we we have to we have to deploy the tools to look at a lot of data very quickly.

Gerardo Dada

39:41 - 39:45

Great. Thank you.

Howard, anything you want to add to that?

Howard Holton

39:45 - 40:49

I mean, I I I think you nailed it. Right? If you look at human psychology, just keep this short, humans only have the ability to keep about 10 things in their head at any one time.

Anything beyond that is is a little bit overwhelming. So much so that numbers larger than 10 are sort of in a different place in the brain than the numbers one through 10.

Right? And so when we look at AI and we look at automation and we look at kind of maturity and evolution, you really have to get the basics down. And then as the complexity and specifically the nuance that you're looking for starts to take place, that's where AI and automation become really critical.

Right? AI can take into consideration thousands of dimensions when making a decision. Humans, like I said, can take in less than 10.

Well, when when most of the decisions and most of the improvement you can you can make to your environment can be made on one or two or three or four dimensions, you don't need AI. It's gonna complicate things.

So what you really need to do is get the basics down, really develop this operating model, and then use AI to improve and mature the operating model, not start with AI and hope it's gonna work.

Gerardo Dada

40:49 - 42:22

Right. And then I will add one thing.

You guys remember, man, it sounds so cliche, but it's still true even in an AI world. At the end of the day, garbage in, garbage out, guys.

I don't care what what do you call AI where you put lipstick on it or you put, like, a a toupee on it, it's still garbage in, garbage out. Right? AI is not going to fix itself.

So I think it's very important part of all these monitoring strategies and x, y, and z strategies. There is a data cleansing aspect that is extremely important.

Otherwise, you're just going to train your thing on on on on on bad data, and you still want to have bad data or bad outcomes at the end of the day. Right? So I think there is if people are looking for a magic bullet or magic solution, that still doesn't exist.

So there is still a lot of work that needs to be done in understanding the infrastructure, understanding your ecosystem, understanding the layers of things. Do I have the right data? Is it mapped the right way? And then AI and all these other things can can improve.

I mean, we've been doing AI for for I I remember buying a system that came out of Israel in in the early early two thousand called Smarts. I'm aging myself, but Smarts was like this root cause analysis, great company that VMC bought.

But the same thing, garbage in garbage out, right, too. So we had to feed a lot of where to feed it constantly good data for us to get good outcomes, and this hasn't this is not going to change.

Gerardo Dada

42:22 - 42:46

So so speaking about feeding the right data to the system, if you're gonna be mature and consolidating all this, you you need to have this data about all these dependencies, right, about all the systems and data and stack that we're talking about. And that's where Internet performance monitoring and IPM concerns.

So how does IPM fit in this broader observability space in this context of modernization and maturity?

Gerardo Dada

42:46 - 42:49

You broke up, Gerardo. Do you mind asking that? Sorry again.

Gerardo Dada

42:49 - 43:01

Yeah. The question is, how does IPM, Internet performance monitoring Yeah.

Fits into the broader sort of ability space in the context of consolidation, maturity, modernization?

Gerardo Dada

43:01 - 44:41

I think, again, I'm a big believer that in order for somebody to paint a picture of what's going on, you need different signals. Right? So, you know, there is not a single signal that gives you everything.

You need multiple signals. The signals needs to be cleaned, and it's a it's a signal rich, data rich environment we have to live in, right, in order for you to run your business better.

And so, the data that, for example, we we focus on is how can we give you Internet level data. How is the Internet doing? Because if you think about it, like, let's say I'm running an ecommerce company, and suddenly my CEO calls me.

It's like, hey, Mehdi or Gerardo or Howard, why are sales down in Germany? Right? Okay. Well, did we have a power loss in Germany? Did like, what happened the other day in Spain? Did we have did did was there a catastrophic issue? Was there a power outage? Was there a cloud outage? Was there so you have to go through these layers of triangulation and figure out exactly what was the root cause.

And so our job is to give as much of those signals to those companies to be able to triangulate as quickly as possible and eliminate the the potential root causes. Was there a CDN outage? Was there a network outage? Was there catastrophic even worse? Right? And by the way, every second counts in these things.

Like, most of our customers lose millions of dollars on a per minute outage kind of thing. And so, going back, it's just it's it's very important that people need to understand that you need to collect as many signals as possible, make sure they are clean.

So, therefore, you can answer those questions. But it starts with the business, like, what are we trying to do here?

Gerardo Dada

44:41 - 44:44

Great. Howard, anything you want to add to that answer?

Howard Holton

44:44 - 45:41

Yeah. Yeah.

I would say that the the beauty of IPM is it makes you not myopic. Yes.

Right? Traditional monitoring only monitors the things that are right in front of me, and so I'm kinda myopic. Right? I I would choose a full view of a narrower space than I would choose a wide view and and a really and be really shallow.

Right? I would rather have a really deep view of the four or five applications that mean the most to my business that are really shallow view and, thus, blindness over the 500 that I actually have. And I think that's a diametric change in the strategy that we use to determine what does success look like.

Well, success looks like successful use of our core applications by the core users, and sometimes those users are not part of our organization. Right? And our ecommerce users not they're not part of our organization.

And so I'd really rather have a deep view of those things that really truly matter than a than a a very shallow view of a bunch of things that don't matter so much to the business.

Gerardo Dada

45:41 - 45:43

That's great. Thank you.

Gerardo Dada

45:43 - 45:50

I'm actually days of, the the days of, like, why I couldn't reproduce it, but my computer is over.

Gerardo Dada

45:50 - 46:13

Yep. We have one question from the audience, Mehdi, our our friend asking this with the question regarding, I would like to get more insights on the statement.

When everything was on premises, we got full control. When I look back, who really did a good job in these times? I think this whole on prem and full control is one of those nostalgia moments people have.

Who wants to comment on that?

Howard Holton

46:13 - 47:19

Well, I completely agree. We tend to always put on rose tinted glasses when we're looking to the past.

Right? But but, oftentimes, the other thing people are saying is it was more simple. Right? When it was physical hardware and physical servers and pre virtualization, forget about cloud, it was far more simple.

And so when we look back, what we're really doing is is we're kinda making a complexity statement when we look back with those rose tinted glasses. Because like I said, right, you can only keep 10 things in your mind.

I used to have hundreds of IP addresses memorized. I couldn't do that today.

There's no way I could do that today, much less would that be the right way to do it. Right? So I think I think we just kinda look back, and and we have this idea that things were more simple, that we we could wrap our arms around more and feel kinda comfortable like a cozy little weighted blanket that we knew what all this stuff was.

And and today, it's just simply not there. So so I do think there's a little bit of rose tinted glasses.

I do think, you know, we probably didn't have as as as good or easy control as as we believe that we did. But it doesn't really change the perception either.

Right? Nor does it change the.

Gerardo Dada

47:19 - 47:20

fact that.

Howard Holton

47:20 - 47:22

you need to do better.

Gerardo Dada

47:22 - 48:41

Yeah. In general, there was just less moving parts.

Right? I mean, ultimately, this goes back to this thing. They they were just they're still moving parts.

But I remember, like, I was monitoring our corporate email system back in the days. It was simple.

We had an exchange server cluster running in in the Tenth Floor. We control the routers, the switches, the m all the employees were in the the same building.

You know, the we had three, four ISPs coming in. So that's the or or the only unknown was literally the three, four ISPs.

But other than that, it was fairly under control. It there were less moving parts.

There were less external moving parts. Today, you have to worry about your starlink connection.

Right? I mean, how do you monitor that? We have customers that have starlink on the roofs of their of their of their offices. They want to monitor that.

So now you're in the satellite Internet, satellite space monitoring. And so they were just simpler times.

Right? You had a much simpler set of prerogatives. But, again, it goes back to one thing.

We were okay if if the Exchange server would go down for an hour. That's true.

That's ultimately the main thing. Nobody would be screaming if email was down for an hour.

Today, nobody will tolerate a fifteen second blip.

Gerardo Dada

48:41 - 49:34

That's true. Great.

So so that, takes us to the end of the the conversation. So thank you, Howard.

And and and maybe I would like to ask the audience to, participate in this fall. I'm gonna open the poll and give a few minutes, or maybe a few seconds to respond.

There's only a few questions, and this is basically about if there's anything we can do to to help you. And in the I'll remind people that there's also a docs tab here where you can actually see and download the maturity model from GigaOm and some other white papers.

And so if, I would appreciate if you if you can finish your call, it will help us serve your needs. And with that, I'll ask maybe start with with Howard.

Any final thoughts, any any last recommendations in this concept of maturity and in observability?

Howard Holton

49:34 - 50:09

Yeah. I would say, you can't monitor what you can't measure, and that includes your own strategy and your own maturity.

The document's free. The it's really designed around the end user.

It's not designed around, the kind of the vendor view. So just take a look at it.

Just try to see where are you and where do you wanna go and what are the steps that are involved in doing it. The the ability to to simplify this and the ability to talk in a clear way about your own strategy and your own goals is immeasurable in value.

And that's really what we designed the tool for. So I hope everybody gets some good value from it.

Gerardo Dada

50:09 - 50:10

Great. And that's great.

And and.

Gerardo Dada

50:10 - 51:07

you you you you you guys are going to think that Howard and I, conspired, but, actually, we haven't. I was going to say what gets measured gets improved.

Right? I I I truly believe in that. I think what gets in what gets measured gets improved.

What you measure, you find out. Thanks.

You you go and fix that. And then at the end of the day, I would strongly encourage all the IT leaders, everybody that is involved that that attended today, you have a space that's running a better business.

Don't run away from it. IT folks usually are shy, and they run away.

I built a career on trying to sit at the table and improving my business by giving data to people to improve their business. So I think we you have an incredible role and to play in running your businesses better.

This is not just an IT thing. This is running businesses better.

Thank you so much for attending, everyone.

Gerardo Dada

51:07 - 51:19

Thank you for your time, Mehdi, and thank you, Howard, for being our guest. We appreciate all your insights.

And, thank you, everybody, for joining us today. Have a great day, and, thank you for the reminder to download the resources available.