Making Technology Performance Sexy
Technology performance. Reliability. Scalability. Resilience.
Is this not the sexiest, most fascinating, and iconic topic that is on top of everyone’s most “talked about” list as we enter 2012?
OK, maybe not.
In fact, it may be that in a market prone to hype and deals over drinks, technology performance may be considered dull, unimportant, and inconsequential. That is, until the very technology you depend on:
- Is fraught with error and downtime that directly degrades revenue
- Is hard to scale so that you need to tap into capex or allocate resources to the technology, versus focusing resources on growth
- Is fragile, quirky, unreliable and your provider is far too willing to consider fragility “normal business practice."
Technology matters. The strength, completeness, and reliability of technology is inseparable from business performance. This may, indeed, be the sleeper story of 2012 and beyond.
So let’s bring it into focus. Our CTO, Jim Butler, who is a key architect of OpenRTB 1.0 and soon to be released OpenRTB 2.0, talks to a number of technology issues that are intrinsic to business performance. First up is technology reliability as measured by up-time.
“Is the platform up?” A simple and fair question, but the answer often depends on who you ask. “Really?” Caveats abound not just in technology, but everywhere you turn. Ever actually read the fine print at the end of a car commercial describing the lease terms? Of course not; you’d need a high-def DVR, intravenous caffeine, and time to burn. Drug caveats are worse, unless you like dry-mouth, eye pain, and in rare cases death.
“So is the platform up?” Your service provider may claim an impressive 99.99% uptime or even the 99.999% gold standard. Typically applied on a monthly basis, these numbers roughly allow no more than about 5 minutes or 30 seconds of downtime per month, respectively. But grab some coffee and read the fine print, or better yet just ask. Many providers exclude scheduled maintenance from the equation, which might include software deployments, network updates, hardware preventive maintenance, etc. These are all valid tasks, but if they’re offline during the process, you need to know; their “real” uptime percentage may have a few less 9’s. Some less reputable providers may even classify an operational state as uptime when the platform is minimally capable of responding, but with major features unavailable – some that may be critical to your business.
I advocate a customer centric view of platform availability. If our customers can use our platform normally with the requisite performance and the features they count on, then we’re up. Otherwise, we’re not up, irrespective of how valid our reasons may be. I take this position not only because I’m proud that we’ve built a platform that enables high “real” availability, but because we too are customers relying on the platforms of many other companies. But one advantage of sitting where we sit in the mobile ecosystem is that we can shield from our customers the downtime of other platforms by aggregating multiple media providers.
And yes, our platform is up (no caveats to follow).