When I advise leaders on a strategy that includes data science, I ask them to consider the probability that their great idea won’t bear fruit. It’s a tough space for visionary leaders to enter — their optimism is what makes them great visionaries. That said, most data science ventures don’t turn out, and most leaders aren’t in touch with the reality that the odds are against them. Having a fallback plan makes good sense, and having a fallback plan for your fallback plan makes great sense.

For instance, when I rolled out an upgraded loyalty platform for a large financial transaction processing company in 2010, we built four plans that successively addressed the failed execution of its predecessor plan. Fortunately, we never had to pull the trigger on even the first fallback plan; however, we were fully prepared for any and all scenarios. It’s a prudent approach that I recommend for you as well, because data science is a risky endeavor.

The colors of cautious management

The best leaders have a backup plan for their backup plan. In fact, when running a strategy that incorporates big data analytics, I suggest you have a series of colored plans: green, yellow, red, and blood red (or black).

  • Green is your plan of record. It’s not overly optimistic but rather a realistic plan that accounts for a reasonable amount of risk. (This is also known as a happy path or a gold plan.)
  • Yellow is a contingent plan. It satisfies your minimum expectations, though it may not be optimal or ideal.
  • Red doesn’t meet your minimum expectations, but it doesn’t set you strategically backward either. This is the do-no-harm strategy, where you’ll probably sink cost and time, but you’ll maintain your competitive standing in the marketplace.
  • Blood red is your worst case scenario. Nobody wants to be here, but you must plan for it nonetheless.

Green and yellow tacks

The first hurdle to overcome is an emotional one, and then it’s time to get to brass tacks (or rather, green and yellow tacks in this case). Setting up a series of fallback plans starts the same way you develop any execution plan: with a green plan. Work with your analytic managers to structure a plan that makes sense. Take into consideration what you’re trying to accomplish, when you need to accomplish it, and the team of resources that will get you there. You should also address the human aspect of team development, quality control and assurance, and most importantly risk.

Your risk management plan should incorporate all the major uncertainties, including positive uncertainty (or opportunity). You could develop a gold plan from here to take full advantage of a particularly favorable execution; however, what’s more important is that you consider a point in your risk management plan where you reset (i.e., lower) your expectations. This is the beginning of your yellow plan.

Your yellow plan vs. your green plan

There are some important differences between your yellow plan and your green plan. First, your yellow plan’s target is a minimum outcome, not an ideal outcome. For instance, you may set out to build a breakthrough product or a service offering, and have a fallback position of creating something that’s only distinctive.

Second, your yellow plan is a derivative of your green plan, so it’s not necessary to recreate everything that went into your green plan. You may need to make some tweaks in your quality plan and address the transition in your team development plan, but most of your execution components should be in place with your green plan.

Finally, there is a very clear decision point (with gating criteria) at which your green plan is abandoned for your yellow plan — this must be as clear-cut and unemotional as possible. Many leaders make an emotional decision to abandon their yellow plan gate when the time comes and execution isn’t going well; this is obviously a bad idea.

Your remaining fallback plans are created in the exact same way: Your red plan is a fallback for your yellow plan, and your blood red plan is a fallback for your red plan.


Smart leaders are visionary and pragmatic. Unfortunately, the dark side of pragmatism is considering undesirable outcomes.

Corporate strategies that involve big data analytics carry extreme risk, even with the best team of data scientists. Make sure your green plan is supported by a yellow plan, your yellow plan is supported by a red plan, and your red plan is supported by a blood red plan.

If you plan to walk on a tightrope, it’s good to have a balancing pole, better to have a safety net, and best to have a big balloon under the safety net.