Jiri Knesl

Posted on 11th June 2020

ICE score for decision making

news-paper Management | News |

At Flexiana, when we are deciding on improvements, we use a couple of ways to decide and prioritize. It can be a strategic objective or directly related to it. Or it can be a huge issue we have to solve. And then there are ideas for improvement that come naturally, even they are within our scope.

For these ideas, we use ICE scoring.

So what is it?

  • I means Impact
  • C is Certainty
  • E is for Effort

$$ICE Score = {Impact \times Certainty \over Effort}$$

Impact is expected income from given activity. 1 stands for $1000 per year (or in total if possible to estimate). These $1000 can come from new income, from savings, or indirectly. For example $10,000 in potential deals might generate $1000 of real income.

$$Impact = {Revenue|Savings \over \$1000}$$

Certainty is how sure we are we will get the value. At this moment, we use a scale from 0 to 1. We multiply how much we believe it will happen times how much we have outcome under control times how big part of expenses we can save (minimum is 0.1).

$$Certainty = SuccessEstimate \times OurControl \times min(ExpensesRecoverable, 0.1)$$

SuccessEstimate and OurControl are on scale 0-1, ExpensesRecoverable are on scale 0.1-1. Average item we implement has average Certainty 0.42, average Certainty overall is 0.36.

Effort is calculated as expense where 1 is $1000 spent per year on people, licenses, or other expenses.

$$Effort = {Expenses \over \$1000}$$

All this leads to interesting conclusion:

$$ICEScore = {{{Revenue|Savings \over \$1000} \times Certainty} \over {Expenses \over \$1000}}$$

Which after simplification leads to:

$$ICEScore = {{Revenue|Savings \times Certainty} \over Expenses}$$


$$ICEScore = {{Revenue|Savings \over Expenses} \times Certainty}$$


$$ROI = {Revenue|Savings \over Expenses}$$

$$Risk = {1 \over Certainty}$$


$$ICEScore = {ROI \over Risk}$$

So at Flexiana we use ICE score as some kind of risk-aware ROI. And to invest into some initiative, it should have positive ROI with risks accounted in.

This approach promotes:

  • Activities with lower commitment
  • Activities which we believe will work
  • Activities where we have the result under control

With this approach, we ended up with:

  • Paying for SaaS as a monthly commitment for the first few months
  • Building our own systems
  • Doing smaller experiments before committing to larger ones
options of decisions No, Yes, Maybe

Internally we have some tables and default measures for estimation. So 1 average new customer is worth X revenue. An Internal employee working fulltime on a given agenda is worth Y expenses, etc. With that, we are able to calculate ICE score for tons of ideas in under a minute.

Then, in our Notion, we have a backlog of ideas sorted by ICE score. Once we have an empty slot for new change (more about this in a different article), we pull the one that’s the highest and evaluate it more. Then, we probably proceed with the implementation.

We don’t delete ideas that have a low ICE score, because the impact of some things scales with company size but the expense is basically the same (typically some SaaS we pay on a per user basis).

At this moment I am very happy to have ICE scoring implemented because it is not only a tool that pushes us to think about potential revenues and expenses, but also about risks and at the same time, revenues and expenses are inputs for our budgeting & project success evaluation.

Although Flexiana is already established, profitable and growing on our scale, in most cases we focuse on only one thing with big expenses & large risks at a time. On top of that, we usually work on multiple improvements that are quite cheap and safe.