How to monetize your data assets successfully
A manual for turning data into products
The debate about surveillance capitalism and data privacy has led to data trading being perceived as dirty business. It may well be for personal data, however these concerns cloud the ambitions of owners of non-personal data to monetize their data assets.
In this post I will outline the major enablers of data monetization and give practical guidelines how to assess data and their product market fit and how to commercialize them.
Data in this context refer to non-personal data which are being created by way of doing business (or research), through processes and transactions that can involve people, objects and/or machines. These data are a side product of primary activities, sometimes referred to as “exhaust data”.
Five drivers increase the value of data
01
The volume of data continues to double every 3 years. This trend is accelerating further, driven by digitization of all processes and life in general. Smart devices, IoT and 5G and other drivers play their role, leading to an estimated 1.7 MB of data per second being generated for every person on earth by 2020.
02
Data analytics capabilities have increased over the years and you can roughly divide the corporate world into two camps: the ones who get it, i.e. use insights from their own data to drive efficiency, sales, customer and employee satisfaction; and the ones who don’t get it. The lack of technology, business processes, talent and culture sets both camps apart and need to be solved to remain competitive.
03
Data exchange hast long suffered from a number of issues around provenance and trust. Blockchain addresses this by enabling all data transactions to be fully traceable and transparent along all steps — and creates a difference by storing transaction records in a tamper-evident way. Of course, other issues such as data interoperability or lack of appropriate data remain, but trust into origin and journeys of data sets will make a massive difference.
04
With Artificial Intelligence and Machine Learning being adopted broadly comes a hunger for training data. The reason that facial recognition techniques are so advanced lies — partly — in the availability of sufficient and diverse training data. With every competitive company becoming an AI company, this hunger will grow. “Data is the new algorithm” is a new mantra, with AI training data creation emerging as a new business model. More generally, data can improve core processes, allow development of (or be) entirely new products and fuel new business models such as subscriptions to pretty much all machines in the world.
05
The emergence of a new trading infrastructure allows the Buy and Sell side to meet. Data brokers and data marketplaces allow to discover previously unknown buyers and use cases and enable a structured trading environment in which trust, quality and pricing transparency can (!) emerge. Advanced data companies begin to set data acquisition budgets aside and professionalize their procurement processes to quickly assess, buy and deploy new data sets.
Two ways for data products to emerge
When attempting to monetize data assets, there are — as always in product development — two routes to follow: Outside-In and Inside-Out.
For outside-in innovation, you start with discovering what your customers need (market need) and then assess whether you can serve this need (product-market fit). While this is often the desired approach, it has some challenges in the space of data monetization, given that you likely not yet know your customers.
Potential buyers of data
So, who could be your buyers, what is your market? In this dynamic environment, there are currently three groups of players standing out — financial services, companies applying machine learning and data marketplaces.
Financial services have long been good at collecting and reading clues to assess investments and opportunities, ideally earlier and exclusively, to grant a competitive edge. Applying “alternative data” (altdata) is far advanced and big moves such as the acquisition of one of the leading providers of alternative data, Quandl, by NASDAQ or Bloomberg announcing alt data integration have shaped the industry. An entire ecosystem for data trading exists, including services for matchmaking, pricing, etc. In a reverse move, the experience with buying and leveraging alternative data sources has led some banks to start trading their data, too.
Examples of investment intelligence being derived from uncommon data sources include using
#
location data of people to assess activities in production facilities (think oil refineries or mobile phone production)
#
satellite images of parking lots to assess traffic at retail stores
#
public registration data to crosscheck a car manufacturer’s claims
And it is not just VC, PE and banks spending money on alternative data, but public agencies are getting started as well. The UK National Office of Statistics has announced to source scanner data from retailers to aid prediction of inflation rates, citing “improved product coverage, high frequency of collection, as well as potential cost savings” as possible benefits.
Data is the new algorithm
Machine learning is changing all industries and McDonald’s acquiring an AI company was just another reminder that every company will become an AI company. AI and ML are only as smart as the data that trained them, with training data having become a bottleneck already.
Some companies have started selling their exhaust data to ML-applications and have found this so lucrative that they have starting creating training data de-novo, which shows how the data industry has gone full circle from the deliberate creation of data for selling — to selling exhaust data — to create data for selling, informed by the former.
Data marketplaces as major changemakers
Data marketplaces are a means to find new buyers which didn’t exist before. If you remember what you did with the stuff in your crowded basement before the arrival of Ebay (leave it, throw it out or sell to 20 people at a yard sale) and look how this has changed today, you can draw conclusions on the potential of data marketplaces.
It will turn data into products which never were planned to be sold, it will empower sellers who didn’t know how to reach an audience, and allow buyers to tap into previously unavailable supply. Add to this the typical marketplace infrastructure around trust, transactions and pricing transparency. There are a number of marketplaces emerging, with the IOTA Data Marketplace one of the most prominent examples. In parallel, we see the old economy diversifying into this space (e.g. Continental and HPE for data from autonomous cars).
Profiling your data assets
For inside-out product development, you assess what you have (supply) and then find a buyer in the second step. The value of data assets is being based on both objective as well as subjective factors. Profiling your data (data audit) means to inventorize your data following a number of dimensions to assess the potential value and be able to pitch your data product in the best way.
Accuracy, History and Timeliness are dimensions that are relatively easy to describe and to assess, even in the absence of knowing who your market may be and what it needs. Data accuracy depends on the representation of reality or use of a trusted source. Data history looks at the available periods of data including possible repetitive measurements in your data set for internal cross-checks. Data timeliness talks about the frequency with which data are being captured and were captured over time.
Data value is highly subjective
Completeness, Accessibility and Enhancements are easy to describe but hard to assess when not yet knowing users and use cases. Data completeness checks both on whether all fields are available and whether all instances are covered (at the right granularity). Data accessibility describes the available data formats and languages. Data enhancements describe the current state of data in terms of value add such as normalization, standardization, categorization and other elements.
Finally, there are fully subjective factors that determine the value of your data and these are relevance (is the data applicable?), objectivity (is the data neutral?) and scarcity (how difficult is it to get this data through other channels?).
Commercializing successfully
Once you get a rough understanding of possible buyers and have audited your data offering, the commercialization process can begin. Let’s look at three essential elements of particular relevance to getting good business out of data.
Sales process for data
In this emerging market there is (certainly outside Finance) no established selling infrastructure or process. The challenge starts with understanding who your data user may be and what need / pain point your data can support with.
If you think finding users will be hard: finding buyers will be harder. Value creation from (external) data does often neither (yet) have an owner nor a budget. It requires a resilient and disciplined business development professional to help identify the right people who have the power and budget to acquire data.
Pricing your data assets
When selling to a new market, finding the right balance between explaining the value you think your data has, and grasping the value it really has, is another important task. This process of consultative selling will allow you to figure out the right price.
As exhaust data don’t have a cost-tag and there is also no market price yet, data transactions require value-based pricing which you can only get right through listening carefully. Data value is not intrinsically defined but depends on the highly specific user cases for interpreting this data.
When quoting a price, don’t have your engineers quote, but quote your engineers. They often get through peer-to-peer conversations a better understanding of the use cases and possible value add than the business folks being wary of exchanging information that will be used in price negotiations.
When quoting, keep two more things in mind:
#
data ingestion and transfer often don’t run as smoothly as envisioned and can require resources for set-up and data onboarding
#
data interpretation doesn’t often run as smoothly as hoped and the buyers will require support and chaperoning in reaping the benefits. This turns many data companies (temporarily) into data consultancies — assure that your pricing model covers that.
Business models for data monetization
The third elements of successful commercialization is a fitting business model. In this emerging market, there are a number of business models emerging, with no clear runner-up (yet). One-off data deliveries vs ongoing subscription access, access to data vs to results, access to live vs historical data are just three aspects to consider, with plenty of combinations possible.
Consulting services and custom data work are additional options, for which the consultative sell has hopefully clarified the need or desire.
IP can be a headache
If you are still reading then chances are good that the novelty of the still emerging market dynamics doesn’t scare you but rather increases your appetite in making new business by monetizing data assets. A last work on intellectual property before you get started.
I recommend that you discuss three questions with your legal department:
01
Do you have the rights to sell the data? Questions such as who and what helped to create the data often lead in the right direction.
02
Is there privacy at stake? Do you have or need consent from individuals and/or would data anonymization approaches solve any concerns?
03
What rights to grant to create a sustainable, ideally repetitive business from data? Use vs resell rights, rights to data vs results, one time vs multiple case are some of the questions to consider, tying in closely with the business model options discussed earlier.
Will this get big?
Estimates on the possible market sizes are rare and everyone likes to reference the commercial scale of the oil industry (which also took years to develop a way to trade oil, by the way). One specific estimate looked at data from IoT devices:
“Based on initial projections, by 2030, blockchain-enabled IoT data marketplace revenue could reach $4.4 billion. The market value of the data being transacted via these exchanges could rise to $3.6 trillion by 2030 — by which time, more than 1 million organizations would be monetizing their IoT data assets and more than 12 exabytes of data would be transacted every day.”
Comments explicitly invited! Where do you see data monetization happening and emerging, what are your favorite learnings?
