Performance Metrics and Data Paralysis
Collection doesn’t equal detection. When your eyeballs are pulsing from the stress of looking at multiple screens, logs, performance metrics, and messages of interpretation from your team; it’s challenging not to experience your own personal system overload.
“Thinking too much leads to paralysis by analysis. It’s important to think things through, but don’t use thinking as a means of avoiding action!”
– Robert Herjavec
Data is an undeniable aperture into operations and interactions but it can be challenging to funnel your data effectively. Partly because there’s so much of it, partly because it can change so quickly.
How do you keep your wits and harvest the actionable information you need from a constantly expanding void of information?
Falling into the Trap of Data Paralysis
Water, water, everywhere … data may not quench your thirst but you can certainly drown in it.
Multiple data sources means you have the potential to paint an encapsulated picture of your business as determined by your online presence. Visitor demographics, mouse clicks, applications … There are many avenues to monitor; and just as many to manage so your strategy shouldn’t be, collect everything first and deal with it after.
In a perfect environment there would be an option to view data omnisciently on a single screen but as mere mortals, consolidation is the next best thing. Consolidating your data increases your level of manageability, the more you’re willing to invest in your monitoring the more options there are to add reports, subaccounts, and users to delegate and act on the data you collect.
An efficient way to get a consolidated view of your data is to work with a provider that generates your data through one platform, and with a vendor capable of integrating with other providers. and applying your existing data.
With data you want to be able to drill down and drill up. Website traffic is a good example where a change in perspective can reveal much more about how users see the site.
Consider RUM monitoring, in which drilling down into real user session data can reveal load times, user locations, browser information, etc. to answer, what is the singular user doing? What is their experience?
Starting from one user’s experience and drilling up shifts the interpretation of your aggregated data – is the user experiencing an error? Is that same error affecting other users in the same way? The curse of data is that one small insight can shift your perspective of the whole; giving you deeper insights into your monitoring and subsequent alerts.
You can’t watch every data point 100% of the time and you certainly don’t have time to read alerts for every minor fluctuation. Instead, set alerts to monitor variations. If your checks are maintaining the status quo, there’s no need to commit yourself to a constant state of observational awareness.
Effective monitoring can be set up with escalations to alert the appropriate people based on changes; if a check goes from up to down, if it stays down for a certain amount of time, if you get a massive influx in traffic, etc.
Who Needs to Use Your Data?
Like water, data is a necessity that continuously sustains your business; making data not just a primary resource for DevOps and your operators, but also for your marketing team, sales, and support staff. That’s a lot of people that need varying levels of access to different datasets and outputs.
A simple way to streamline the presentation of specific data is with an Internal Status Page; like a live billboard, a Status Page can centralize current data in a snap for your team but as we’ve said, sometimes it’s beneficial to be able to drill a little deeper.
At the Enterprise level especially, an account may have hundreds of checks, and several users that need to take a closer look; status pages are great for alerting but they don’t double as portals for DevOps and engineers. Multiple user accounts at varying levels of permission boost productivity by providing access and control for key members of your team.
What are Performance Metrics
Performance metrics, and Key Performance Indicators (KPIs), reflect and log a single type of data. As an example, take a SaaS company required to report SLA accountability. reporting metrics and uptime percentages is a data-fueled necessity. SLA Reports can display the composite data from specific checks and components, and are one way to consolidate data in a clear way. When it comes to reporting data, there are two types we should consider.
Vanity Metrics vs Actionable Metrics
Vanity metrics have two uses; user reassurance/trustability, and sales. Vanity metrics are displays of data that are attractive to your users but not critical in terms of understanding your performance.
Actionable metrics are data that have the power to move the needle on your growth and development meter.
Displaying both helps build reputation as long as the actionable metrics are clear.
Managing Data & Performance Metrics in Real Time
It’s one thing to sift through logs of data when time isn’t pressurizing your efforts. When solving an incident, data can be your best friend, but only if you can call up the right points in an instant.
You want to isolate key metrics from key data points. Intentionally gather data in a way that can be used throughout your company as KPIs. An example of this is a DNS check, or checks setup to provide detailed records on DNS records and servers and escalations for optimizing alerts.
Energizing Your Performance Metrics
This is the Tortoise and the Hare approach. Metrics determine management; what you look at influences your decision making so if you can’t benefit from analyzing it, what’s the point?
Any manager in any sector will tell you that organization equals speed. So while an audit of your data isn’t something you want to take on in the crux of an issue, it’s the foundational analysis that will expedite your decision making and data access when speed matters.
You’ve gotta be thorough and go slow in order to be fast in your response time. An audit of your data and examination of how you’re using it is a must for expediting results.
It may sound counterintuitive but the best way to overcome data paralysis is to make sure you have enough data but also that you have the infrastructure to support it. Data after all, is like water; a source of power if you direct it.
Minute-by-minute Uptime checks.
Start your 21-day free trial with no credit card required at Uptime.com.