Architecting Data Driven Marketing Experiences

Architecting Data-driven
Marketing Experiences:From Insights to Action


In our hyper-connected, socially networked marketplace, consumer expectations have assumed a critical importance. This is especially true for the millennials and Gen Z consumers who are expected to comprise 63.5% of the estimated 7.7 billion global population by 2019. Moreover, studies indicate that at least 85% of millennials prefer to research online before making a purchase, seeking to maintain the upper hand in the consumer-enterprise relationship.

Now, let’s examine what is happening in the industry and how it is responding to this shift. According to an IAB report, more than 91% of marketing leaders believe companies take decisions based on customer data, but 39% of the marketers can’t turn this data into actionable insights. Couple this with the findings of a Bank of America report, and you will start to witness a pattern. The report states that computational costs have declined by 27% in the past decade and are expected to further decline by another 22% in the next decade, which will provide a thrust to digital decision making. In fact, Forrester states that insights-driven leaders will steal $1.2 trillion from data-aware firms between 2018 and 2020! The potential is near to infinite for digital transformation.


Digital Innovation TransformationDigital is the next stage of the technological evolution, the previous one being information and communication technologies

Now, let’s take a closer look at how consumer behavior has evolved over the years.

In the brick and mortar model, customer engagement used to be one-to-one, mostly representatives in a retail store or officials in a bank branch interacting with customers from behind a counter. Customers did not have access to information. The cost of market research was also high in those times, which explains why consumer expectations were so low. Moreover, the data generated was insufficient and technology penetration low.

With the advances in information and communication technologies, customer relationships became one-to-many. In addition to the brick and mortar offerings, mass broadcasts of emails started taking place, with the hope that the customer receiving those emails may need their products and services, even when they may not. Also, centralized contact centers were set up to offer product and services over phone. Websites were chiefly informative, allowing zero to limited sets of transactions, relegating marketing to a monolog.

With the advent of digital, however, people started to interact with intelligent technology through mobile shopping apps and websites. Personalization of offerings became sought after due to a steep rise in consumer expectation, which arose from easy access to information. Now, switching costs are at an all-time low—how much does it cost to switch from Walmart to Amazon? And hence, there is no such thing as a loyal customer anymore. Marketing is increasingly becoming all about staying relevant to the customers when they are in need.

So how do you—a marketer in today’s intensely competitive times—address these consumer tendencies? For you, availability of data is not a problem. The advent of new technologies brings in its wake all types of consumer data—behavioral, transactional, and individual identity–related. The pace at which this data gets generated is also phenomenally high! To deal with such data, the skill sets required are niche and expensive, and the number of resources required is high. And this is why advanced analytics techniques are gaining notable importance and becoming increasingly mainstream in the marketing space. And, this is exactly where DecisionMinesTM fits in. Its data science and machine learning algorithms ingest the data and help you derive predictive and prescriptive insights, taking away all the number crunching from you and democratizing decision making. This is especially relevant in a scenario where the application of data analytics for decision making is at the embryonic stage and organizations are still struggling to make sense of their data.


Now, let’s try to understand the challenges faced by the marketing community in light of these evolving paradigms.

Meet Tina James. Tina is a Chief Marketing Officer. She is responsible for overseeing the planning, development, and execution of her organization’s marketing and sales initiatives. During her day-to-day tasks, she often faces roadblocks in terms of the following issues:

  • How can I maximize the ROI across campaigns based on their performance to ensure higher conversions?
  • How can I maximize the ROI of a campaign by optimizing the media mix?
  • How can I sell my products at higher profits?
  • How can I effectively identify demand patterns and ensure future readiness?
  • How can I clear my non-performing products inventory?

Tina has access to top-drawer web analytics systems, CRMs, marketing automation platforms, and DMPs, which generate large amounts of data. However, her team is not able to make data-driven decisions on a daily basis and she strongly feels that their performance is below par.

Let’s see how DecisionMinesTM leverages data from various technology platforms through various stages of the marketing and sales funnel and translates it into measurable actions. At each of the marketing and sales funnel stages, data is generated from the aforementioned systems. For awareness, it is chiefly Ad trafficking and operations systems. For the consideration stage, it is usually the web analytics system and product catalogs. Abandoned carts and wish-list data is usually analyzed for drawing conversion stage insights. And reviews, service CRMs, and referral engines are leveraged for retention and advocacy.

Raw data from these systems is fed into DecisionMinesTM to figure out what has happened. For example, Facebook responded better for product A. Or a set of customers (mostly premium) always buy the latest products.

Next, diagnostic insights are drawn from the identified events to know why something has happened. For example, Facebook audience demographics were more suited for Product A. Or premium customers buy latest products to stay abreast of the latest trends in the market.

From the new data that is generated on an ongoing basis, patterns similar to historical patterns are identified in their early stages to flag up a warning or an unidentified opportunity. For example: I want to launch a campaign for a product similar to Product A. What’s the optimal budget for an awareness campaign and which channels should I target? Or Product B is getting more traction and should not be discounted.

A prescription set is prepared based on the identified patterns and then turned into actionable insights using a prescriptive dashboard. These insights prompt decision makers to make informed actions based on the available data and not on individual, subjective outlook.


Now that we have the right outlook, let’s put your marketing dollars to work.

In the aforementioned scenario, Tina has invested USD 500 in TV marketing, USD 200 in Facebook marketing, and USD 300 in YouTube marketing. This means, 30% of her total investments are in YouTube, 20% in Facebook, and 50% in TV.

Decision Point Unoptimized Media Mix AllocationUnoptimized media mix allocation
Decision Point Optimized Media Mix AllocationOptimized media mix allocation

The DecisionMinesTM Campaign and Media Mix Optimization solution enables you to do all this, but more efficiently, through prescriptive dashboards.

DecisionMinesTM has the capability to fine-tune the investments as per demand units and conversion rates from the organization’s data reserves to indicate where to allocate the investments and elevate the marketing performance.

The DecisionMinesTM Campaign Optimization solution monitors the performance of ongoing campaigns and recommends moving funds from a low-performing campaign to a high-performing one to help you achieve a balance between product awareness and revenue targets.

At a specific campaign level, it monitors channel-level performance and prescribes fund reallocation between low-performing and high-performing channels. So, if acquisition from Facebook is cheaper and a volume game, why spend that money on low-performing TV?

Prescriptions from DecisionMinesTM help you fine-tune your expenditure across campaigns and media channels, depending on your role.

Customer Buying PreferencesBridging the gap between customer buying preferences and the optimal offer

Let’s proceed to the next use case called the ‘next-best offer’. Next-best offer is a form of prescriptive analytics that helps marketers and their organizations judge customer spending habits better and guide their marketing efforts towards connecting with customers to close a deal.

Here, on the x-axis, we have the marketing funnel, and on the left y-axis, we have customer buying preferences in terms of pricing, product category selection, among others; and on the right Y-axis, we have the seller’s value proposition.

Let’s assume a customer searches for a pair of shoes. The customer engages with the platform, applies several filters, adds a product to the bucket, but doesn’t buy it. This product falls outside the customer’s buying preferences—let’s say, chiefly in terms of price.

Since the customer added the product to the cart and abandoned it, the customer is a qualified lead—what should be done now?

We can squeeze in some marginal popular products that can be tailored to the buyer’s preference and also pick popular products that fit the customer’s buying preferences, and send these options as the next-best offer.

The next-best offer enables you to motivate customers to move from one stage of the marketing and sales funnel to the next stage. Just imagine the concept as a magnet attracting customers through various stages by keeping them engaged and channeling their efforts to the best-fit deal. Based on the customer’s engagement pattern—whether the customer has applied various filters, or searched for the latest and premium products or discounted products—the customer can be classified as a premium or thrifty customer, or a free delivery hunter.

Next, basis the behavioral and product purchase data, products can be re-priced as per popularity and classified as premium, standard or low-performing products. High-performing premium products will be offered at higher prices or zero to lower discounts to chiefly target the premium customer segment. Underperforming products will be offered at reasonable discounts to chiefly thrifty segments, allowing for win-win selling.


Measurement is a discipline, and its systematic use is a source of progress. Measuring is a matter of understanding the past and present and then it is also projecting oneself into the future so as to steadily advance. Measuring is also about identifying the right metrics because having too much information kills information, and not having enough information is also a problem.

Being data-aware is the first step towards progress. But with data getting generated continuously and at a huge pace in today’s turbulent business environment, quick decision making is required for effective control. Anyone with first-hand experience of working with insights-driven organizations would agree that they act way faster through prescriptive dashboards than their data-driven cousins, earning better ROI through effective identification of risks and opportunities in their early stages, taking the right actions, and also pushing business users to think about their business in new ways.

Like what you’re reading?

Get the PDF version of this page now!