How to get your machine learning project approved

10 Nov 2021

Machine learning projects are high profile, but the tech is still new and still evolving. That makes business decision makers nervous. If you have a machine learning project idea here are three tips to get your machine learning project approved.

  1. Avoid confusing the audience — don’t talk about the technology.
  2. Increase interest by make the business outcomes clear.
  3. Reduce fear buy saying how much effort is involved.

If your idea pitch can do these three things you have a much better chance of getting approval.

1. Avoid confusing the audience, don’t talk about the technology

This may sound counter intuitive but you don’t need to talk about technology to get a technology project approved.

You may be a machine learning expert, but your audience almost certainly isn’t. If you talk about how the ML works you will confuse and lose the focus of you audiences. And if the business audience doesn’t understand what you are asking for they will not approve it.

The purpose of your proposal is not to educate the business on how the technology works. It is to convince them of the benefits the technology brings and that benefit is worth funding. You can talk about benefits without talking about how the technology works. To do that you should focus on the business outcomes.

2. Increase interest by make the business outcomes clear

Business stakeholders care about business outcomes. When people care about something they are automatically interested in hearing about it. When pitching your machine learning project make sure you state the business outcomes. Tell them what it will it do for the business.

Business outcomes fall into three main categories:

Financial outcomes: Financial outcomes focus on costs and earnings. To give your ML idea the best chance of approval include the answers to these questions in your pitch:

  • Will the project increase earnings for the business?
  • Will the project reduce costs?

Time outcomes: Everything a business does takes time. Increasing efficiency and getting more done in less time are valuable business outcomes. To define the main time benefits ask yourself these questions:

  • Will existing processes happen faster?
  • Does the capacity increase? Can the company do more in the same amount of time?

Quality outcomes: Improved quality at the right price is beneficial for a business. Quality can be measured in the quality of performance, outcome, or experience. Every business and every team have different quality outcomes they care about. Think about the type of quality your target business stakeholder cares about and then answer these questions about it:

  • Does the experience for the user, customer, or employee improve?
  • Is the performance of a tool or process improved?
  • Is the outcome better?

Machine learning projects tend to deliver time and quality outcomes. These then lead to positive financial outcomes.

Common business outcomes from machine learning projects include:

  • More accurate reporting to support decision making
  • Identifying trends to inform product development or market strategy
  • Automating manual data processes
  • Predicting customer behaviour in specific circumstances

If any of these things are valuable to your business then include them in your pitch.

Describe what your machine learning project using these three categories and the business stakeholders will be interested in what you have to say.

3. Reduce fear by saying how much effort is involved

New technology is an unknown and unknown means uncertainty. Uncertainty is risky and business doesn’t like putting money into risky projects. Therefore, your pitch must reduce the fear of the unknown.

The two main types of fear are fear of cost and the fear of failure. If you reference other projects in your pitch you’ll help address these fears. It is unlikely you have other machine learning projects to reference because the tech is so new. Instead, if your company has implemented new technology in the past, mention that. Remind people that new things can be done but only do this if it was a success!

Showing you have thought through the costs for the project also reduces the fear of costs. Make sure you consider the ongoing maintenance cost as well as the build cost. Showing you’ve thought about costs and benefits increases confidence. And more confidence means more chance of approval.

Conclusion

Machine learning projects are new and the business is understandably wary of the risks. If you have a great idea for a machine learning project you can still get the support of wary business owners.

If you want to get your machine learning project approved, stay away from technical language. Instead, focus on the financial time, and quality outcomes the technology will deliver. Make sure you know roughly how much effort the work will take and compare it to the business outcomes. Do this, and  you’ll give your machine learning project the best chance of approval.

Originally published on dzone.com