Learning Analytics and AI Best Practices from Fractal

Groups within companies trying to improve their analytics and AI capabilities can learn from each other, or they can learn from other organizations that provide similar services to their business. One of the latter group is Fractal Analytics, a global analytics and AI services company led jointly from the US and India. Fractal was founded in 2000 and now has more than 3500 employees and 16 branches. It recently received a major investment ($360 million) from private equity firm TPG and is valued at over a billion dollars.

Fractal’s success and growth is an interesting story in itself and an indication of how important analytics and AI have become to large organizations. But after talking with co-CEO Pranay Agarwal, I came to the conclusion that the company offers a lot of lessons to in-company analytics and AI groups. Below are five Fractal attributes that other companies could use.


  1. Focus on decisions and how to improve themFractal’s mission is to empower every decision in their customer businesses. They believe that better decisions mean better results for their customers. This focus on driving decisions means that the company’s approach to solving problems is backwards decisions: they start with the decision to be made and then think about how they can improve it. To empower customers to make better decisions, they include not only the resources of structured and unstructured data, analytics, and AI, but also design thinking and behavioral sciences. They involve different types of decisions, but most of them are fast, repetitive, and with a high level of feedback from data. Any individual decision of this type carries a relatively low risk of substantial loss. They start by looking at the industry they work in, map out the value chain or industry drivers, and go one step further to determine what decisions are needed to improve the drivers. Each of these approaches can also be applied by analytics and AI groups within companies.
  2. Clarity about product range and possibilities—Analytics and AI are very broad areas. Each service provider should specialize in a relatively small set of offers that they can repeat and reuse many times. Fractal is quite a big company now, so it can have a wider range of specialties. As I mentioned above, Fractal focuses on quick and repetitive decisions, including “next best action”, or the next conversation with a customer; what price to charge (dynamic pricing); which channel can best serve customers; predicting supply and demand; managing revenue growth and similar issues. The company has developed a series of platforms that support each of these common types of decisions.
  3. Combining Organic Growth and Acquisitions for Talent and Capabilities—Organic growth is often the best way to maintain a culture, but given the scarcity of analytical talent and specialized capabilities, it can also be beneficial to make targeted acquisitions even if your company is not active in the analytical services business. Fractal has made several acquisitions, including Neal Analytics for cloud offerings, Samya.ai for revenue growth management, and Final Mile, a behavioral science consulting firm. It has also made majority investments in other analytics-related companies, including Analytics Vidhya, a data science community and education company. The combination of organic growth and acquisitions has helped Fractal create a wide range of capabilities.
  4. Build a strong culture with specified values—An analytics or AI team within a company will naturally adopt some of the broader organizational values, but there is also scope for establishing cultural principles within the smaller group. Fractal’s leaders believe that their values ​​have been instrumental in the company’s successful growth for more than 20 years. The four values ​​expressed include putting the customer first (measured by the Net Promoter Score, which remains above 70); learn and grow (with the Fractal Analytics Academy and Analytics Vidhya, and they also design programs for clients); think big and act fast (which they achieve, among other things, by reinvesting 10% of their income); and extreme trust and accountability. The latter value implies that they adopt a positive intention from their customers and colleagues. They have been on the Great Place to Work list for five years and have received that recognition at all five locations this year.
  5. Integrate an ethical orientation and competence—Nowadays everyone seems to realize that analytics and AI have an ethical dimension. It’s too early to do much about it in most companies, but I would say that every analytics/AI organization needs some kind of ethical framework or guidelines, and should have a governance structure in place soon. Not surprisingly, Fractal already has these components. Agarwal told me very clearly: “The people who are driving the adoption of AI in our society need to make sure that it is used ethically. We are driving the adoption of AI, so we have to invest in AI ethics.” They have an internal team that oversees ethical issues and a framework to manage them.The framework allows them to evaluate and score any analytics or AI solution on criteria such as transparency and bias/equality. framework not only internally, they also “produce” it for clients, and every analytics organization should be moving in this direction.

Just as internal supply chain groups can learn valuable lessons from companies that do supply chain work as their primary activity, for example UPS, DHL, and FedEx, analytics and AI groups within companies should view companies like Fractal Analytics as a model for building internal successful analytics and AI practices.

Leave a Comment

Your email address will not be published.