Rebuttal: Five Things I Learned About APM By Watching Marketing Eat Their Own Dog Food

Our Marketing VP, Frank Days, explained in a recent post five things he learned from eating his own dog food.

Essentially, he explained that he uses SharePath to monitor our customer facing websites (As I have mentioned in the past we use SP extensively for internal uses). The key points can be summarized in the following tidbits: get the right data, enable IT Operations, know your baseline, and make page load speed an obsession. Let’s take some time and put an Operations perspective on each of these concepts.

1. Get the right data
My buddy Frank correctly explained the conundrum facing most IT Operations teams: “The sad truth is that most people just don’t have the right data.“

As a typical Operations guy, I spend hours looking at many tools. The DBA has one tool, the network manager has another tool and the server guys a third. When something goes wrong these individual tools may all show a green screen! As I’ve said in the past:

You can have a totally green screen, alerts indicating that the CPU is at 100%, and official monitoring tools telling you that everything is A-OK. But you can’t listen to that. You need to keep going, keep investigating, and keep learning about the environment so that you know for yourself that everything is OK.”

The key is finding the right tool which gives you the right data to prove the performance of your critical applications.

2. Enable IT Operations
As a COO, I lead a group of incredibly hard-working and talented experts who are very adept at wearing many hats. The challenges and surprises we encounter in our customers’ environments can sometimes seem insurmountable. But we always find a way. The main point is we’re constantly trying to juggle several problems, and we don’t have time for problem detection. What Ops guys need is a tool that auto detects the interdependencies in an environment and gives true, up-to-the-minute measurements of service level performance. This empowers IT Ops to analyze performance and dramatically reduce problems.

3. Know your baseline
The key metrics for Marketing and Operations may be different, but to be successful in both, you must understand what standard operations actually look like. Is a 2 second response time good enough for customers? What about 2.1? How can you tell if you haven’t been monitoring it consistently? It’s even more critical for Ops guys. We have key SLAs to maintain. Ops needs to know precisely when 2%, 10%, or I dare say 20% of transactions are not meeting the requirement and to push the “I am so sorry” email in advance for poor performance rather than waiting for someone to call. You must proactively manage these service levels so you can avoid nightmare outage scenarios (usually during weekends).

4. Make page load speed an obsession
After we put our work into improving our page performance we saw a 12% increase in downloads of SharePath Express. Speed is crucial! When you have phones that are running so fast today, people are expecting to see the same fast response time on every website and app page. For us Ops guys, I believe end to end visibility is the obsession. The response time of critical transactions can never be fast enough! You can never be too diligent in measuring KPIs and SLAs. And finally, nobody in the history of IT Operations was fired because they improved a good site to be a better one :-) .

With the growing complexity of IT environments and with the amount of changes we deploy every day, it’s important to take note of what makes your life easier as an Operations professional. If you gather the right data, enable the rest of your Ops team, keep track of your baselines and obsess over your KPIs and SLAs, you will always be heading in the right direction. Even if you have to eat your own dog food along the way.