Too much data and not enough time and resources to leverage it intelligently? Need someone to assist you in decision-making by mining the data you have on your Customers, Products, Employees or Operations? Have an urgent need for sophisticated data analytics? Or just plain data management? Get a 2-pager on us. And contact us .

April 16, 2014

MI Intellisense : An intelligent ordering system for CPG firms

CPG companies have to constantly tackle growing competition and consumer expectations while driving overall profits. To remain profitable, companies look to maximize sales of products to retail outlets in a manner that boosts outlet sales and increases customer footfalls. Secondly, companies look to push out historically slow moving SKU’s from retail outlet’s portfolio and include new innovative fast moving SKU's. These two objectives coupled with the right inventory size help boost sales and profits for both the CPG manufacturer and the retailer while keeping the end Consumer satisfied. Various considerations like past consumer demand, size and channel of outlet, target audience and region of operations go into determining the right product assortment and quantity.

A leading CPG company was facing a similar dilemma on how to boost retail outlet purchases based on SKU recommendations. They understood that a ‘one-fits-all’ approach will not work and pursued a data-based approach.

Based on the stores’ historical sales pattern, product assortment, growth trend, categories sold, sales mix and so on, the outlets were segmented into distinct but homogeneous segments. This segmentation helped tailor the product mix in the store to its relevant customers. Through segmentation, an outlet profile was generated which aided in making decisions on the stocking/de-stocking of product lines and SKUs at an outlet level. This in turn helped boost outlet sales and achieve customer delight. After the outlet profile was developed a recommendation engine as built to identify the best SKU’s and their optimal quantities for each store.

Determining the best SKU’s for each store
Based on the segment of an outlet and the best representative SKU’s for each segment, we identified the SKU’s that will boost outlet sales while keeping intact the manufacturer’s and retailer’s business objective. New innovative SKU’s representative of the outlet segment were also recommended for each outlet to enable customers get access to new and innovative products.

Determining the right quantity for the best SKU’s identified
Along with determining the best SKU’s for an outlet, the right quantity of the SKU had to be determined in order to drive the overall bottom line and balance inventory. Also care was taken to ensure the new SKU’s recommended did not cannibalize  sales of existing successful SKU’s. A non-linear constraint programming approach was used to optimize SKU quantity. Constraints were based on store’s profile and store’s buying capacity.

This process has to be repeated periodically in order to cater to new SKU launches, constantly changing product choices and growing competition.

Below is a snapshot of the SKU recommendation engine developed - MI Intellisense.

When a retail outlet depending on its profile/channel of sales has the right product assortment, intangible & tangible goals like increasing consumer footfalls, increasing consumer loyalty are met. In this way, by cross selling and up-selling the right SKU’s in the right quantity for a store, the CPG company’s objective to sustainably boost sales of their products can be met.

February 27, 2014

Leveraging Text Analytics to Identify Customer Pain Points

Over 80% of data being created today is unstructured; however over 80% of data being analyzed is structured. 

One of the key text data elements that almost all businesses have is Customer feedback data gleaned from various channels – online surveys, emails, forms, etc.  Extracting actionable insights from this data element can help a business across almost all its key focus areas – designing better products, increasing customer satisfaction, loyalty and retention, designing optimal promotions and pricing strategies, etc

Quick service restaurants (QSR) with an array of products and service quality experiences across locations face challenges to analyze latent information hidden in customer feedback surveys.   To circumvent this problem, for a large QSR, we converted customer feedback text into words/phrases and based on their frequency of occurrence, combined them into clusters. A combination of Silhouette Score & Gap Statistic was then used to identify the right number of clusters. Themes of the cluster were understood by looking at word clouds of unigrams, bigrams & trigrams. Based on this clustering exercise, word clouds for various ‘themes’ were generated and analyzed to highlight areas of opportunity and concern.

This exercise allowed us to help the QSR in the following ways: 
  • Identified high priority areas of concern on a weekly basis which if not taken proper care of, might result in customer churn
  • Highlighted features that customers liked & which could be scaled across regions to give Customers a seamless experience
  • Better targeting of discounts and coupons were possible by clustering price-conscious regions
  • Top burning issues identified by region and store, and weekly reports generated and distributed to senior management for corrective action
        More details can be found here .

January 11, 2014

Application of Decision Sciences in Retail Banking

Marketelligent partners with Retail Banks to increase their revenues and margins by making faster and more informed business decisions using our realm of analytics capabilities that pans through the Customer life-cycle.

More details can be found here.  Please contact us for more information.