Gartner Q&A Part Three: Analytics vs. APM

January 2013

In Part Three of APMdigest's exclusive interview, Will Cappelli, Gartner Research VP in Enterprise Management, talks about APM, analytics and mobile performance management.

Start with Part One of APMdigest's Q&A with Gartner's Will Cappelli: Analytics vs. APM

Start with Part Two of APMdigest's Q&A with Gartner's Will Cappelli: Analytics vs. APM

APM: Your December report - Will IT Operations Analytics Platforms Replace APM Suites? - cites mobile performance as a blind spot. What is the problem?

WC: Great question. There are a couple of problems. One is the fact that at the end of the day, if you really want to see what is going on, you need to get something on that platform. The most obvious way of doing that is to instrument the platform directly. Unfortunately, from a social perspective that is almost unachievable. If employers were supplying us with our smartphones and tablets, and we were using them strictly for business purposes, there are technologies that could handle the problem.

But one of big issues is that a user is not going to be that crazy about putting an agent on their personal iPhone or tablet, nor will an IT operations team be that crazy about having to manage and administer that agentry. There are even places in the world where it is illegal to do that. You cannot do that in Germany, for example, without taking all sorts of precautions.

But there are other ways to get at mobile performance information. There is a lot of work being done now with JavaScript injection technologies, so once a webpage renders on a mobile device, the webpage itself start to act as an agent. But then the issue is: How do you interpret that data? How do you correlate that with other data you are gathering?

I don't think it is an insoluble problem. It is a difficult problem. Solutions are still in progress. A lot of money is being invested. It is a huge customer requirement. I believe that through some combination of endpoint instrumentation, made suitably lightweight and painless to manage, and JavaScript injection, we will see the emergence of some kind of economical solution. But we just don't have it yet. I think it is perfectly within the industry to reach there. But we're just not there yet.

APM: I am sure we are going to see many next generation mobile monitoring solutions in 2013. Do you think by this time next year the problem will be solved?

WC: I am purely speculating here, but I think we will see the outlines of the solution. You will see smaller vendors with very attractive solutions and you will see larger vendors putting similar solutions on their roadmaps. If we have this conversation next year, I think there will still be a lot to execute but the vision of what mobile application management is about will be largely settled and complete at that time.

APM: Your other December report – IT Operations Analytics Technology Requires Planning and Training – says that today's IT ops teams need more training to make analytics effective. Is this going to be a new breed of analytics administrator, or is it easy to train the current teams?

WC: I think you will have a new breed. With the Big Data issue there is a lot of talk about the emergence of a "Data Scientist" role. Whether or not that materializes on the business side, I do think you'll see the need for that kind of capability. You will see professionals answering to that role. I don't know what they will be called – they may be called IT Operations Data Scientists – but there will be a new role. I don't think you can take your average IT operations professional, even professionals involved in app development or application operations – their expertise, even if it is profoundly technical, is not technical in this direction.

There are two sources where you can find this expertise in current organizations. One is the capacity planning teams. In order to cope with managing virtual environments they brought in that level of expertise. Another area is the Six Sigma graduates floating around in Global 2000 enterprises, who are often very fluent in some core statistical concepts. I think if you can remove some of the macho martial arts rhetoric around Six Sigma, it can be transposed into a skill set that makes use of these tools.

APM: I would like to wrap up the interview by talking a little bit more about the stats in that first report. I think people could read that and make assumptions about the “death of APM”. I think you and I both agree that is not the case at all.

WC: It is always tough. On the one hand, if you speak vaguely it doesn't tell you anything. On the other hand, if you publish numbers, people will reason as they will.

I am definitely not saying APM is dead. In fact, I think it is going to grow at a healthy clip. Just not the growth we've seen before. There are many reasons for that, and some of it has to do with analytics, but another really big point is what we were just talking about. Part of it does have to do with the fact that one of the outstanding problems for enterprises is this mobile application problem. If the big vendors don't have a solution yet, the customers are not going to make the investment now. It is almost like a direct quote from a number of calls I have had: “We think the technology is great, but right now our problem is mobility, and the product doesn't address that.”

If you and I are both right, and by this time next year the mobile problem begins to be mitigated, in 2014 you could very well see some of the growth that APM lost this year kicking back in.

There will be a part of the APM market that will forever be lost to analytics – in the sense that there will be some problems that in the past may have been seen as being solvable by APM technology that is in fact being solved by analytics technology – but that is par for the course. That is not a negative judgment on APM. And people shouldn't take the report that way. It is just noticing a change, not proclaiming the apocalypse.

APM: I would add that even if certain vendors are not going to experience as high sales numbers because they focus on a particular technology, in general I think analytics just makes APM stronger. We are moving towards a more progressive level of APM.

WC: You could certainly make that case.

Click here to read Part One of APMdigest's Q&A with Gartner's Will Cappelli: Analytics vs. APM

Click here to read Part Two of APMdigest's Q&A with Gartner's Will Cappelli: Analytics vs. APM

ABOUT Will Cappelli

Will Cappelli is a Gartner Research VP in the Enterprise Management area, focusing on automation, event correlation and fault analysis, management system architectures, and real-time infrastructure issues. Prior to Gartner, Cappelli served as director of Research for EMEA at Meta Group and held a variety of analyst and management positions at a number of major research firms, including Forrester/Giga Information Group, Ovum, New Science Associates and Real Decisons Corp.

Related Links:

Gartner Report: Will IT Operations Analytics Platforms Replace APM Suites?

Gartner Report: IT Operations Analytics Technology Requires Planning and Training

Gartner Analyst Profile: Will Cappelli

APMdigest's Interview with Will Cappelli in 2011