Big Data

A Digital Transformation Lesson: Open Source Business Models

Bill Schmarzo By Bill Schmarzo March 29, 2018

The year was 1994 and I had the fortunate opportunity to stumble upon a company – Cygnus Support – that was “selling free software.” I remember telling my mom that I was Vice President of Sales & Marketing of a company that was selling free software. After a very long pause, she replied, “Is your resume up to date?”

Cygnus Support sold support contracts and custom consulting projects for GNU development tools (gcc, g++, gdb) to companies looking to accelerate their time-to-market in the embedded systems market. Our value proposition was very clear and compelling for embedded product customers. We could fix compiler bugs in days, not months, which not only accelerated time-to-market, but also reduced the size of their embedded code by avoiding costly workarounds.

At the time, Cygnus Support executed a new, rarely-seen business model. They leveraged the open source concept to put development tools into the hands of software developers that fast-tracked time-to-value and de-risked product development efforts.  Eventually, Red Hat bought Cygnus Support and validated the open source business model (see Figure 1).

Figure 1: Red Hat Stock Price Performance

Fast forward to today where open source projects are the norm. Hadoop notably launched the open source Big Data market, laying the foundation for open source projects in IOT (Liota, Kafka, Nautilus, EdgeX Foundry), and machine learning and deep learning (TensorFlow, Apache Spark ML, Caffe, Torch).

We only need to observe their rise in popularity to understand more organizations seek an open source model to steer their overall business plan. However, identifying the best, or at least the most optimal, open source strategy remains unclear. To understand the best course of action, let’s go to our old friend, and 18th century economist, Adam Smith for some guidance.

Understanding Adam Smith and Sources of Value Creation

Adam Smith, in his seminal book “The Wealth of Nations,” described value creation in two ways:

  • “Value in Exchange” – defined as the value of an asset based upon how much you can get paid for the asset. Our common accounting practice is based upon the “value in exchange” concept (think about how what you pay for an item determines depreciation schedules).
  • “Value in Use” – defined as the value of an asset based upon how much value you can generate from the use of that asset. This is a discussion where economics is the branch of knowledge concerned with the production, consumption, and transfer of wealth.

Some open source companies, like Red Hat (Linux) and Hortonworks (Hadoop), have adopted the “value in exchange” business model, with the goal of selling support contracts and custom development services. That was the business model that we pioneered at Cygnus Support.

Other open source companies, like Google (TensorFlow) and Facebook (Torch), have embraced open source projects with the goal of leveraging the open source community to improve their respective platforms upon which they create value.

Key point is this: Google and Facebook don’t sell these open source products, but instead use the open source community to expand the capabilities of the platforms upon which they create new sources of value. This “value in use” strategy exploits the dynamics that a community can create a better product faster than the creator (Google, Facebook) can create on their own.

So, from a business model perspective, it’s a value in exchange (Red Hat and Hortonworks) versus value in use (Google and Facebook) decisions. It’s a product (to be sold) versus tool (to be used) decision. It’s a business model decision.

Open Source Business Model Lessons from Google

Google’s TensorFlow business model is based upon getting more developers to embrace and expand the capabilities of TensorFlow faster than Google could do on its own. The article “Reasons Why Google’s Latest AI-TensorFlow is Open Sourced” highlights Google’s open source strategy:

“In order to expedite the evolution of its ML and move towards a robust AI, TensorFlow needs to be exposed to new data sets, some of which might be proprietary data of the company/user that decides to use TensorFlow for applications. Google hopes that once TensorFlow is deployed across applications by different users, these users can then contribute to the original source code of TensorFlow with their upgraded code, as mandated under the Apache APA license. This would aid the company to roll out a [more] comprehensive AI engine in the future.”

Business Models and Value Chain Analysis

Let’s call upon another old-school friend of the Big Data MBA community – Michael Porter – to understand how his Value Chain Analysis technique can help guide our business model and digital transformation discussion.

In my original Strata presentation back in 2012, I shared with the audience how they could use Michael Porter’s classic (Old School) Value Chain Analysis technique to identify where and how to apply big data analytics. The goal was to deliver material financial, operational, and competitive benefits to the organization (see Figure 2).

Figure 2: Michael Porter Value Chain Analysis

You can find the original blog “Big Data MBA: Course 101A – Unit III” here (there were two other MBA techniques that I covered that day and you can find links to those materials in the first paragraph of the blog).

Let’s say that you are in the retail business and looking to leverage big data analytics to “optimize in-store merchandising effectiveness.” Let’s use the Value Chain technique to understand where and how to apply big data analytics.

  • Inbound Logistics: Use real-time Point of Sales (POS) analytics to predict (score) out-of-stock situations and prescribe corrective actions to suppliers and distributors as to what products, in what quantities, to deliver to what stores at what times.
  • Operations: Use real-time POS and RFID data to: predict merchandise demand; forecast slow product sales; prescribe sales and promotional actions to mitigate the impact of slow and non-movers; and optimize in-store / on-site inventory.
  • Outbound Logistics: Integrate and analyze social media with real-time in-store mobile app data and external event data (e.g., a large area event, unplanned construction work on a major travel artery) to identify and quantify merchandising and store traffic trends and model event-driven logistics impacts. Respond by prescribing in-store merchandising actions in order to optimize merchandising insights that impact stock and inventory levels for in-flight campaigns.
  • Sales / Marketing: Use conversion attribution analysis across search, display, mobile, and social media to quantify the online variables that are driving merchandising performance in order to optimize ad placement, keyword bids, and messaging in-real time.
  • Service: Combine social media and POS data with your customer loyalty data to create more-frequent, higher-fidelity customer scores for retention, fraud, up-sell/cross-sell, and net promoter scores that guide customer loyalty programs and promotions.
  • Infrastructure: Deploy predictive, real-time merchandising dashboards that predict in-store and department merchandising problems and prescribe corrective actions.
  • Human Resources: Combine social media data with data from local job sites and competitors’ hiring pages to predict at-risk employees and prescribe retention actions.
  • Technology: Use in-memory analytics to predict merchandising performance problems and prescribe corrective actions via an actionable mobile dashboard app.
  • Procurement: Combine merchandising images and in-store video surveillance data with POS data to quantify in-store partner promotional program effectiveness and negotiate better terms and conditions with key suppliers.

Michael Porter’s Value Chain Analysis provides a framework against which we can make decisions about where and how we can apply our digital assets to derive and drive new sources of value creation.

Digital Transformation is About Business Model Transformations

Digital transformation is not about application development or cloud or even big data analytics. Digital transformation is about leveraging the organization’s newly minted digital assets to derive and drive new sources of value creation.

Digital Transformation is about the integration of digital assets (data, analytics) and capabilities (AppDev) into an organization’s processes, products, and assets to create new sources of value creation – to improve operational efficiency, enhance customer value, manage risk, and uncover new monetization opportunities (see Figure 3).

Digital Transformation starts with an understanding of the organization’s value creation process for if you are not delivering new sources of organizational and customer value, why bother.

Note: I will have a paper forthcoming that details some recent work at the University of San Francisco and the National University of Ireland Galway on the Digital Transformation process. Think Value Chain Analysis meets Economics meets Design Thinking to help guide an organization’s digital transformation process.

Bill Schmarzo

About Bill Schmarzo


Read More

Share this Story
Join the Conversation

Our Team becomes stronger with every person who adds to the conversation. So please join the conversation. Comment on our posts and share!

Leave a Reply

Your email address will not be published. Required fields are marked *