The following data dramatisation was adapted from an actual text by xiax solutions GmbH, an “interdisciplinary research development” company (fancy way to say they are IT rocket scientists) based in Davos, Switzerland.

The example of the service availability scenario on xiax’s website shines a spotlight on some unsettling facts and figures. Facts already well understood by systems engineers and IT infrastructure architects. But to decision-making C-Level execs, responsible for a company’s growth and brand reputation, even startling facts may be easily dismissed under the illusion “We run SAP” or the false assurance of a 5-nine SLA (service level agreement).

My rewrite of the same scenario illustrates how xiax’s facts and figures can be made personal for the C-Level executive by dramatizing a story’s stakes to trigger the action xiax wants those decision empowered execs to take next (contact them sooner not later).
Quick Review of what any message we hear/read subconsciously triggers in the mind:
1. Why should I care?
2. What’s at stake?
3. What is it you want me to do next?
Judge for yourself if this story satisfies the Story ROI Test for you.

How 5-nine (99.999%) Uptime = 60’000 Angry Customers

It’s 11:30 A.M. two days before Christmas. Your shopping list is complete except for two people: your boss and your mother. You’ve waited until now to get their gifts because you’re ordering both presents online from an Internet retailer. You browse to the retailer’s website, find and add the items to your cart, enter shipping info, choose “next day delivery,” and click pay now, congratulating yourself your gift-shopping is finished with 15 minutes to spare (next day delivery ends at 12:00).

The screen refreshes. Instead of seeing “thank you for your purchase”, an “Error in transaction” message appears (error code 445). You frown and refresh the page. Same result: “Error in transaction.”

Now its 11:50. You’re starting to stress. You hit the back button only to find you have to re-enter all your payment details again. Bother! You enter them again.

11:56 You hit SUBMIT, holding your breath and praying the payment will go through. Knowing if it doesn’t, the boss and your mother are not getting the nice gifts you thoughtfully selected for them.
The screen refreshes: “Error in transaction” AAAAHHHHHHH!!!!

12:02, it’s too late – Mom and the boss are not getting packages from you this year. You call the retailers customer service number and give up after waiting on hold 30 minutes listening to Frosty the Snowman loop 14 times.

Now if you’re the resourceful type, you revert to last-minute plan B and telephone flower delivery shops near where your mom and boss live.

Christmas Day: Mom calls and says the bouquet looks nice next to the chrysanthemum delivered the same day from her chiropractor.

When you go back to work in January, the boss says thank you, but suggests using a different florist next time because he found your cut flowers on his doorstep, dead beyond recognition, two days after Christmas when he returned home from spending the holiday with his girlfriend.

Think this could not happen because you have a 5-nine uptime SLA? Think again.

Let’s assume you do have a 5-nine (99.999%) available payment transaction processing system capable of handling 1000 requests per second. On the fateful day in question, an “Act of God” interruption causes a 1-minute downtime, affecting 60’000 customers.

Now suppose 5% of them called customer care as the poor guy in our anecdote did at $6/dial cost (-$18K). And 1% of them switched to another source for their future business with a recovery cost of $50 per client acquisition cost (-$30K). Profit is down -$48,000 before factoring the high probability 5% of them tweet, app chat or tell friends how your brand spoiled their Christmas joy.

A final thought before you get back to work: a 1-minute outage is completely ok to occur within a 5-nine Uptime SLA.
Footnote: xiax founder, Christoph Huber, has graciously given permission for this case story and would like to hear stories from execs or engineers who have experienced scenarios similar to the one featured. Original text link http://www.xiax.com

Recent Posts