Everybody knows the old engineering adage: Good, Cheap, Fast. Pick any two.
Now most of us that answer to bosses understand they actually want all three and, on top of it, want “Cheap” to be “Almost Free” and “Fast” to be “Tomorrow.” Or at least before the next quarterly earnings.
Either the air is awfully thin on the executive floors or they have the best stash. Maybe both. But I digress.
After over thirty years of building marketing technology systems I’ve finally realized that we should be using the Apollo program’s Cost/Reliability/Schedule model. But isn’t that the same as Cheap/Good/Fast? Aren’t those words just synonyms?
Yes and no. Let’s walk through them.
Cost.
Let’s look at the total cost of the thing. Not just the OpEx or CapEx required to acquire the technology. But also the people that will run it. After all, it’s not going to run itself, regardless of what the sticky on the 35th floor tells the executives. You have to bake (no pun intended) all of that into the structure, including apparently mundane things like the individual contributors who are going to create and maintain the data pipelines.
Reliability.
It’s not just about “good.” You have to decide how much uptime you want. For most marketing systems two or three nines is enough to run your campaigns. If you want five nines, you’re talking telco-level reliability and megabuck outlays. In Apollo, reliability was critical. If the thing didn’t do what it was supposed to do on the way to the moon, there was no truck roll or emergency phone bridge call that could fix it.
Schedule.
And it’s not about “fast.” There may be good reason to pick vendor A for one part of your system and vendor B for the second part of the system. And vendor A and vendor B may claim a lot of past integrations and successful deployments with customers. Even if true, that schedule just went way longer than you might imagine. Even the best PMO (Project Management Office) staffed by the best PMPs is going to look their Gannt chart explode in terms of dependencies and critical path items. If your company has go fever, your cost just went through the roof and you still might have to yield on reliability.
My recommendation? As much as possible, figure out your Cost, Reliability, and Schedule parameters as early as possible. Yes, your initial plans and timelines will eventually go out the window, but you’ll have a better framework for where to give–maybe it’s going to two nines from three nines for your source level reporting dashboard.
This post is also available at Linkedin here.
PilipBlog 
You must be logged in to post a comment.