Data Monetization Strategy Playbook

10 steps to kickstart your portfolio enhancement journey

In my previous post, I covered the three key imperatives for business leaders looking to transform their enterprise portfolios and grow revenues in this fourth industrial revolution. One of those imperatives is Data Analysis.

If you are a Chief Product Officer or in an equivalent position, what pointers would you give your product management team to help them craft an enterprise data monetization strategy and a credible business case to drive investment decisions? Today, I’ll give you my ten-step roadmap summarized in the infographic below. Read on for the details.

Data monetization is a lucrative endeavor although a tricky one to take on. Both vendors and customers can gain a wealth of operational, performance and fiscal benefits from data analysis, trend mining and predictive algorithms. But to get to that end-game, vendors must navigate an ever-evolving, complicated minefield of regulatory policy and vertical-specific requirements. This is much like a game of snakes and ladders!

In many companies, data monetization is an evolutionary strategy with the potential to impact two distinct areas: Cost and Revenue.

A data compilation initiative initiated to improve internal decision-making – whether it is engineering prioritization, sales campaigns, or usage monitoring – can deliver cost reductions and improve profitability.

But a deeper data monetization strategy, one that I discuss in this article, aspires to become more of a revenue-generating endeavor based on product improvements, increased adoption and/or the addition of new profitable lines of business. Here are examples of this more strategic play:

  • Boost existing product/service adoption through recommendation and benchmarking services. Just look at the rewards that data, history and deep learning algorithms have delivered for Amazon or Netflix in recent years.
  • Deliver analytics-based services highlighting key trends. Offer a paid-for service based on metrics (ex., capacity utilization) which delivers predictive and prescriptive recommendations, or enable security monitoring for operations highlighting out-of-the-norm events.
  • Add new features and capabilities. Think what products like Nest have done for the consumer space or what Splunk has delivered for enterprise-grade products. The potential for data-based services and products is tremendous today in just about any vertical market, be it finance, healthcare, automotives, education, or legal to name a few.

STEP 1: Thoroughly research all relevant policies, regulations, guidelines and industry trends.

Don’t just limit yourself to the minimum set of factors. This is an evolving space and the implications of missteps are enormous!

Let’s take a look at government regulations. Even though the EU rolled out GDPR only a few months ago, chances are you’ve had to accept numerous cookie notification on EU websites since the rollout. That’s the mind-bending pace of regulatory change today in the realm of privacy. Here, in the U.S., the Government Accountability Office (GAO) is mulling comprehensive legislation on internet privacy.

Aside from government regulations, your specific vertical markets and data types may require that you build guard rails as part of your data monetization strategy. Auto manufacturers, for example, have instituted a set of Automotive Consumer Privacy Protection Principles through their automotive alliance. And the FCC has rolled out Broadband Consumer Privacy Rules. These are just two examples of evolving data privacy considerations, one being “principles” and the other being “rules”. Mind these nuances as you define your guardrails.

You should also pay special attention to countries that are blocking data flows across their  borders. The graphic below is a point-in-time summarized view.

For more on this topic, see the Information technology and ITF Foundation site.

STEP 2: Define your underlying data strategy, as broadly as possible.

Formulate your data strategy in collaboration with your Chief Data Officer or CDO (if you have one), your Legal team, the Chief Information Security Officer, the CIO, and the Chief Strategy Officer. Your data strategy is the foundation of the house you are building plus the underpinning of your data monetization efforts well into the future. This foundation should support both today’s plan and any remodeling you may want to down the road. Take time on this very crucial task to avoid rework.

In this step, you will need to make key decisions around governance and compliance requirements across the data lifecycle. How will new datasets and definitions be added to the data lake(s) and who will add them? How do you plan to manage data ownership and access, both for raw bytes as collected and the analysis results?

Your data strategy will be influenced heavily by your specific industry and the geographies it operates in today and plans to in future. Avoid the temptation to solidify your data strategy simply based on technology choices or architectures. The data architecture is an important but small piece of the overall data governance effort. Decisions on data strategy are as much about people, processes and markets as they are about architecture and technology choices.

STEP 3: Build a cohesive data architecture in partnership with your CDO and CIO.

The list of data fields to collect, how to collect them, and how long you want to retain them are all aspects that will take shape and iterate as you build the overall data solution architecture and evolve your build capabilities over time.

Some questions you’ll run into include whether to store raw data or only processed information, where to archive the data, and how easy it will be to recall it in the future. Data security and integrity should be front and center criteria for your data architecture. These considerations will become especially relevant, for example, when you need a way for customers to obtain their data if they decide not to renew your analytics offering or cease to be a customer.

Have a plan from the get-go for connecting your CRM, ERP and Customer Support data sources with the entire solution lifecycle in mind. Beyond connecting the silos, make sure you have the right interfaces to suit your function and/or purpose. For example, most embedded systems will need a REST API, while your data scientists and engineers may have specialized data architectures or database needs to enable optimized processing for AI/ML algorithms.

While this is not a cloud or AI/ML paper, we are all aware of the need and importance today of cloud-based solutions and the scalability and efficacy that an AI/ML-based solution can provide over simple statistical analysis. Take the time to educate yourself on the technologies. Check out this series of videos which I found very informative on AI/ ML from Josh Gordon at Google and from AWS Invent.

You will need a full-stack cloud architect on hand to help evaluate your architectural choices and tradeoffs every step of the way, from inception to first customer ship to a full-scale rollout. Build, buy, or partner choices will be key here. Once you find your sweet spot, engage with your data architect and scientist to build the architectural guidelines and tune the data analysis algorithms.

STEP 4: Work with your Architect and Data Scientist to build a scalable network, compute and storage architecture

In today’s world of cloud services, you do not have to invest in creating the entire Big Data infrastructure on your own. But you will have to balance flexibility, elasticity and the ability to scale with the relative costs of data movement, storage and the relative price you can charge for your solution to cover some of those costs.

Data policy, residency and sovereignty requirements will also impact your infrastructure architectural choices. Your overall business strategy and theaters of operation (both today and in the future) will determine the specific in-country data policy requirements that apply to your enterprise.

Take Australia for example. The country enacted a Personally Controlled Electronic Health Records Act in 2012 which requires personal health records to be stored within the country. If you are a U.S. or European managed service provider looking to assume hospital IT operations in Australia, you may not be able to leverage your current architecture and infrastructure without modifications in Australia.

Beyond country requirements, the top three things to consider in this step include security, scalability and performance.

STEP 5:  Dialog with internal / external customers and partners to list and prioritize use cases for data monetization or reduce costs.

Establish a streamlined and simple method to prioritize use cases, similar to how you prioritize features for your product or service portfolio. This is easier said than done. Think of the connected car example. Consumers are open to sharing some data about their car in return for useful features (think relative-value to relative-benefit ratio), but the willingness to pay will vary based on the use case and your business model. Here are three use cases with varying relative weights from a consumer perspective.

  • Generate proactive warnings to the customer to bring the car in for service
  • Provide safety information and notification to a parent who has a child driver
  • Collect data on driving habits which can be shared with and monetized by an insurance company to offer customer benefits or to charge penalties

The trick is to whittle down your prioritization criteria to the handful that best validate your data strategy and maximize your business case.

In the end, remember, this may be a new area of experimentation for your company, or perhaps your industry! I keep this quote in mind when prioritizing “pioneering capability” development:

If you make all your decisions based on opportunity cost and the fear of failure, you’re almost certain to fail… Protecting against the downside and being conservative in the face of a priority list means that you’ll choose the obvious and the predictable instead of the subtle or the remarkable.

Seth Godin

STEP 6:  Invest in Visualization tools – build or buy.

It is a boon for the data monetization-minded among us that we have numerous visualization options available today. You can use programmable options like Google Charts or invest in a more sophisticated and scalable tool like Tableau. Here’s something to remember:

The idea is to go from numbers to information to understanding.

Hans Rosling

The objective of your choice of graphs is to convey data-driven insights to all key influencers and create a clear understanding of the significance of the data. Create visualizations that best illustrate your most important takeaways for strategic and operational decision-making.

STEP 7:  Develop your GTM / business model.

My preferred method to develop the business model is to start building a Business Model Canvas across use cases, customer segments and channels.

Size your market, identify your priority segments and incubate. Remember that the same use case can be developed using multiple business models.

Spend your time in the hypothesis and analysis phase and your execution phase will go a lot smoother!

STEP 8:  Co-create, validate and iterate your user experience.

Data monetization in business is not about developing a one-time product or service. In most industries, this is still a developing area and you should be prepared to not know all the answers. Here are some famous words from a Chairman of Daimler:

We know that 10 years from now, this industry will be totally different: we’ll have some of the same competitors, and we’ll have a number of totally new competitors. If we continue to do what we did so well, we’ll be toast.

Dieter Zetsche

You want to keep moving like water and heed that warning!

STEP 9:  Fill the data lake as soon as possible and as comprehensively as possible.

A robust data set is essential when it comes to training your algorithms and identifying and validating your analysis and trends.

Sometimes your training data may not be sufficient to build the necessary intelligence into your neural network. This is where GitHub can become your best friend. You can leverage it just as much as your architect or software designer does. Check out some great public datasets here.

With some data types, specifically around object detection and computer vision, you can augment your training dataset using various techniques. You don’t have to be an expert in data modeling algorithms or neural networks to do this. You only have to be knowledgeable enough to ask how much of a dataset is needed to train and develop an AI model to achieve a decent level of accuracy.

Similarly, in areas of support and operations automation, your ability to provide value from the data will depend on the adoption of the solution and vice versa. Take the example of NetApp. AutoSupport capabilities were intrinsic to the product and it helped the vendor achieve high adoption rates for this capability across their installed base. That meant reams of data from current customers which generated high value from the solution in a self-propagating manner. But AutoSupport is not a differentiating feature today. It is used now by a slew of tech giants including Apple, Microsoft, HP, Juniper, and Cisco.

True to the spirit of change and evolution being the constant, the strategic differentiator today for support automation is how much knowledge you can mine from these old datasets. This knowledge can potentially help you understand usage, analyze potential errors and trends, and stratify the results by customer type or customer configuration, so you can provide best practices and benchmarks and connect them to business outcomes more effectively. This is the new North Star for support automation and managed services.

STEP 10: Each step of the way, ensure you have one or more executive champions and sponsorship.

Remember that developing a data monetization strategy is a long-haul marathon. Your company and leadership must be geared for the journey before you go too far! It is bound to be a long road of experimentation and discovery before you are ready for a big bang.

It is wise to set realistic expectations, take baby steps, and spend the time in the hypothesis and experimental stages before you go for a expansive and possibly expensive launch.

According to Gartner, by 2020, 50% of data and analytics leaders will have successfully created a narrative that links financial objectives to data and analytics. Which 50% do you want to be in when the time comes?

Closing Thoughts

In summary, your data monetization strategy will involve policy and guideline development, identifying future-proofed technology and architecture choices, and creating business models aimed squarely at growth and profitability. Perhaps each of these warrants a playbook of its own!

The most valuable lesson you will learn from this journey is the power of experimentation, the experience you gain from all the small experiments and baby steps that it takes to iteratively build a data monetization business.

Over the coming months, I plan to dig deeper into each of these aspects and spark a conversation on effective data monetization. Stay tuned and please share your comments on this playbook!

© 2019 Nupur Thakur. All rights reserved.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

Powered by WordPress.com.

Up ↑

%d bloggers like this: