As the flood of customer information continues to pour in through an ever increasing number of digital touch points, big data use cases for sales and marketing have grown exponentially. The growth of predictive analytics has, in turn, also been driven by customer-focused use cases. Companies are attempting to understand the underlying correlation / causation in those data lakes and turn that information into better targeting, promotions and demand forecasts.
But how can sales and marketing leverage the power of prescriptive analytics to maximize the return of their investments in big data and predictive?
To find out, I interviewed three River Logic veterans about how they have personally helped companies in retail, CPG, shipping and elsewhere apply River Logic’s Enterprise Optimizer® to sales and marketing. Here’s what I learned.
Leveraging Prescriptive Analytics for Sales and Marketing
Trade Promotion Optimization (TPO)
In CPG (or FMCG), EO is used to optimize trade promotion campaigns. We’ve even partnered with a company, AFS Technologies, who leverages EO within their product offering.
Trade promotions are typically done on a weekly basis looking a year out. EO helps determine which campaigns to run, and for which products, by various dimensions including retailers and channels.
EO considers different objective functions like revenue, profit and volume, all depending on the strategic goals and lifecycle stage of a specific product. For example, a company may have products that are cash cows where profit is the focus. Or it may have new product introductions where volume and market share are the key objectives.
Business constraints include promotional budgets (hard number or live accrual), min/max frequency, min pre/post promotional gap, black out weeks, must have weeks, products to promote/do not promote together, forward buys, etc.
A TPO-like model can also be applied to retailers where instead of looking within a manufacturer, it optimizes the target promotions within a category. Of course, with the power and flexibility of EO, both manufacturer and retailer could be modeled together which would make for interesting game theory.
Product Assortment Optimization
In retail, EO has been used to support product assortment optimization. In one example, a telecom provider looked at the assortment of products in their retail stores. EO helps their business users determine the optimal configuration of premium, high end, medium and low price tablets and phones that maximize the total value to the company.
The model considers the cost of the devices, baseline demand, substitution effects (i.e. if I have Product A, what does it do to demand for Product B? or conversely if I don’t have Product A, how many customers buy other products vs. leave the store?), sales associate training and other costs per product line, subsidies, average lifetime contribution from phone plans, min/max and total number of handsets per category.
It also supports portfolio management, i.e. when to introduce a new phone, eliminate old ones, etc. This is akin to almost any category in retail, except that it can be more complicated in some ways because of the lifetime value and subsidies. Produce in a grocery store might be one more complicated example as it decays more quickly.
Associated with assortment but not quite the same, price optimization can be used to determine the optimal baseline price for an item within a category. The detailed unit costing and opportunity values in EO give clear indicators of potential sources of missed opportunity and how to set prices to assure profitability. Elasticity curves by category be considered similar to the assortment example, including how these curves shift when prices of products that can be considered substitutes change.
Marketing Mix Optimization
Similarly, EO can be used for marketing mix modeling. For example, optimizing the mix of paid advertising across channels such as Google Adwords, Facebook, Twitter, LinkedIn, into different audiences, with different messages. Since these channels each report on cost per click / impression over time, that data can be leveraged in a manufacturing paradigm as a resource with a specific variable cost and yield (clicks) over time. A good marketer will understand how those clicks translate into revenue which ultimately allows EO to optimize the spend to maximize that revenue.
Customer (and Vendor) Contract Negotiation
The same analysis done for producing the baseline product pricing can be leveraged for contract negotiation both with customers and suppliers. EO’s detailed unit costing and opportunity values can be applied to more than just products or resources. These very powerful capabilities can be done by supplier, by customer, by channel, by region, etc.
When examined by customer, EO users have found that the true cost-to-serve some of their most strategic customers resulted in limited or negative profitability. Customer rationalization analysis becomes a very powerful contract negotiation tool. The same analysis when applied to vendors often helps to generate more win-win negotiations by identifying terms that would have the highest impact on company profitability but with a low impact on supplier cost.
Why River Logic?
For each of these use cases there are software companies that have created packaged applications to address these challenges, so why leverage Enterprise Optimizer?
Package apps support pin-point solutions very well with lots of predefined inputs, interfaces, reports, etc. However, if you are looking to these capabilities as a system of innovation or differentiation for your company, where you are trying to get ahead and outperform the competition, you should consider the limitations of packaged apps, namely:
- You can’t build your own innovation into packaged apps – but must instead give your ideas back to the vendor or systems integrator. Once they put it into the product it will also be available to all your competitors.
- Related to the above, you must rely on the vendor roadmap or spend a lot of money and time with systems integrator. Many projects fail when companies try to configure packaged apps into things they weren’t originally designed to do.
- There are limited economies of scale with a packaged app as you need one for every silo problem, so it’s ultimately less efficient, and in smaller companies it will be difficult for users to develop a deep expertise on every packaged app.
- It is really important to consider that an initial problem representation may eventually prove inadequate and the solution must be expanded to include other parts of the business, which often proves difficult with a packaged app built-for-purpose. For example, in CPG, companies may want to extend the model to include inventory positions, cost of goods or production. In retail, a company might want to expand promotions or pricing to reflect supplier contracts, seasonality, or special discounting.
In contrast, EO models can be configured to support the exact company need, supporting multiple types of problems as well or better than packaged apps. The company has all the ability in the world to control their own innovation roadmap and gain economies of scale without waiting on a vendor roadmap or the systems integrator to change or customize any code.
Extra credit! Strategic Marketing & Competitive Analysis
If the company can produce the data, EO can be used for competitive analysis within markets. This is a very complex analysis requiring a certain level of sophistication. That being said, River Logic's customers have been able to build a single model of their business and their competitors' business — along with the markets they play in — and run war game scenarios, anticipating how the other competitors will react to their moves in the market.