Grupo Boticário has grown into the world’s largest perfumery and cosmetics franchiser by helping its customers look good—which means it always needs to keep the hottest new products in stock. IBM Analytics helps the group understand what customers want, before they even know they want it—enabling smarter sales, marketing and production planning.

Business Challenge

Cosmetics are an easy way to tap into the latest trends—but what if that new lipstick sells out before you can buy it? To keep its customers in style, Grupo Boticário must predict demand accurately.

Transformation

Predictive analytics is helping Grupo Boticário gain a deeper understanding of what consumers want, before they even know they want it—enabling smarter sales, marketing and production planning.

Results
  • 20% increase in accuracy of demand forecasts over traditional approaches
  • Balances stock and service levels of the most desirable products, boosting sales
  • Enhances corporate agility by enabling real-time insights into customer demand
Business Challenge Story
Innovation is the lifeblood of the cosmetics and fragrance industries: consumers are always looking for the next hot product. But preferences are always changing, so how can the companies behind the products make sure that they release hit after hit?

Donald Neumann, Demand Manager at Grupo Boticário, responds: “The key to continued success in the ever-changing world of cosmetics and fragrance is predicting consumer demand. But because the drivers behind demand are so complex, this is much easier said than done.

“For example, a television commercial might strike a bigger chord with consumers than we anticipated, leading to a spike in demand. Or—and this is a true story—a famous actress wearing one of our lipsticks can result in it selling five times as much as we expected.”

If Grupo Boticário is not prepared for the sudden peaks in demand that result when its marketing activities resonate with consumers, its franchises may suddenly find themselves out of stock of the products that customers demand—leading to a loss of sales and customer goodwill.

Solving this problem is not easy. The lead times for manufacturing cosmetics products range from weeks to months, so it is often not possible to simply “make more” when the supply of a product runs out. Accurate demand forecasting is vital in order to plan production and marketing campaigns months in advance—but predicting the demand for a new product that has not yet reached the market (and therefore doesn’t have any sales history) is a difficult challenge.

To complicate the situation further, due to the volatility of the cosmetics sector, Grupo Boticário does not operate on regular seasonal sales cycles throughout the year: it needs to be able to launch new products whenever fashion dictates. Understanding the effect of seasonality on demand is therefore even more difficult than usual.

“Most demand forecasting solutions need data from regular sales cycles to produce useful results; something we simply do not have,” comments Donald Neumann. “Every year, the total number of sales cycles we have changes, so although we can still see a seasonal influence on sales—for example, we make 20 percent of our sales in the lead-up to Christmas—we can’t do a simple month-to-month comparison between years using time-series approaches.

“Moreover, with the simple models that we previously used to predict demand, it was not feasible to separate the influence of our franchisees’ buying behavior from our end-customers’ real needs and desires.”

Grupo Boticário has more than doubled in size over the last few years, so it has clearly been highly successful in spite of these challenges. However, this growth has also brought with it increased complexity, and to sustain expansion, it needed to continue out-pacing its competitors and delighting its consumers.

“Everything we do—from which products we choose to manufacture to how we market them—comes down to whether someone wants to buy what we’re selling,” sums up Donald Neumann. “So by getting to the bottom of what drives demand for our consumers, we can efficiently deliver the right products through the right channels at just the right times.

“But in order to build the predictive models that would help us achieve these aims, we first needed a better handle on our sales data. With so many lines of business, channels, and franchisees, collecting and consolidating this information was something that we knew we could do better.”

Transformation Story
Finding a more sophisticated approach to forecasting

Aware that its complex business environment made demand forecasting a unique challenge, Grupo Boticário dived into a 14-month selection process for a new solution. The group narrowed down its initial list of potential solutions from seven to two, and initiated proof of concept exercises with IBM and another major vendor of integrated planning software.

“The solution we were looking for would be the first of its kind, so it was vital to us that we could prove that it would work before we made the investment,” says Donald Neumann. “We chose the IBM solution over the competing option because it was much more flexible; this was a key requirement to help us adapt the technology to our business model. Also, IBM offered us an analytics landscape that could be managed by the business, rather than IT. We wanted to give the ownership of the solution to the people who are actually using it: our analysts and line-of-business teams.”

By taking IBM® Planning Analytics (formerly known as IBM Cognos® TM1®), which provides a fast, in-memory analytics engine for financial and operational planning, and combining it with IBM SPSS® Modeler and R solutions for advanced predictive modeling, the combined project team developed a solution that solves the demand forecasting challenge.

Using the solution, Grupo Boticário’s analysts can log into IBM Planning Analytics and view the sales and product data collected from the company’s operational systems. When they want to create a new forecast, they simply press a button to execute a forecasting stream in IBM SPSS Modeler, which is further enhanced with R scripts. Grupo Boticário uses IBM Planning Analytics to prepare the data, which is displayed online for analysts to view using their planning cockpits. Then, the prepared data flows into IBM SPSS via the company’s IBM DB2® database software. Finally, the SPSS solution creates a new model in real-time using a sophisticated custom forecasting algorithm, which the company’s data scientists have developed using the R statistical programming language, and the results are fed back into IBM Planning Analytics.

This forecast is based on modeling the combined effects of known demand drivers, such as advertising and marketing campaigns, pricing and discounts, seasonality and sales cycles, as well as market data from external sources.

Once the forecast has been loaded back into IBM Planning Analytics, the analysts can apply their own unique business knowledge to adjust the results where necessary. This enables them to take additional factors into account—for example, the impact of a completely unpredictable event such as an unexpected celebrity endorsement, which cannot be factored into the model.

Donald Neumann describes: “Our analysts now have the tools they need to isolate demand drivers and create more effective forecasts. For example, they can identify whether price discounts or channel dynamics play a bigger role in creating demand for a specific product, and they can continually adjust and refine their forecasts as new information comes in.

“We are also able to distinguish between the buying behaviors of our franchisees and our end-customers, which gives us a much clearer view of the real demand. And we can even incorporate macroeconomic indicators such as GDP forecasts to see how the wider economy is expected to affect our sales. Furthermore, the solution is IBM Watson® ready! Meaning that we can introduce Watson cognitive intelligence to the process without major changes for the business analysts.”

Results Story
Fulfilling customers’ desires

Even though adoption of the new solution is still in its early stages, impressive results are already on the horizon for Grupo Boticário, putting the group in a strong position to defend and extend its market share.

“Using IBM Analytics solutions, we expect to lift the accuracy of our forecasts by 20 percent compared to traditional time-series approaches,” explains Donald Neumann. “By enabling us to align demand from end-customers and franchisees more precisely, the software will help us minimize stock-out situations without the need to inflate our inventory levels across the supply chain. The result will be higher sales and lower costs, a win-win situation that helps us stay ahead of our competitors. We can also pass on these insights to our franchisees so that the benefits ripple out across our entire supply network.”

By basing its marketing plans on a more accurate understanding of consumer demand, Grupo Boticário can achieve better results.

“We have already noticed in meetings that our analysts have started to speak about customer demand and behavior, rather than the demand and behavior of franchisees—so we’re getting much closer to the real audience we want to address.

“Through the data-driven decision-making enabled by IBM Planning Analytics, we can design marketing plans that bring more people to stores to buy our products. As a result, we will put our marketing spend to the best possible use for maximum returns.”

With a better understanding of ever-evolving consumer demand patterns, Grupo Boticário is finding the recipe for prolonged success: fulfilling the desires that consumers don’t even know they have.

Donald Neumann concludes: “Cosmetics and fragrance purchasing is highly emotional, and IBM Analytics solutions are helping us understand the myriad factors that make consumers choose one product over another. By putting these insights to work, we can ensure that we are the ones bringing beauty to buyers’ lives, rather than our competitors.”

View more client stories or learn more about IBM Planning Analytics.