Data to dollars: Thoughts on valuing data as an institutional asset

A core principle of data governance is that data, regards of who creates or has custody of it, is an asset of the institution. But what’s an asset? Some ground-breaking work in Canada and Europe might point the way to putting an actual dollar value on our data.

For the first time, Statistics Canada last year tried to estimate the value of the country’s stock of data. The total value of data, databases and software, and data science in Canada is between CAD $157 and $218 billion, the agency estimates. (1)

This is a first stab. Like every national statistical organization in the world, Stats Can is playing catchup with the explosion of digital data capture and processing and its importance to the economy. Data is clearly part of a country’s productive capacity, but it doesn’t show up in national accounting statistics such as gross domestic product (GDP).

The same for universities. Buildings are tangible assets recorded on the balance sheet. Data is intangible and, unfortunately, invisible to generally accepted accounting principles. What if the university were able to value data as a capital asset?

The value of software and analysis is more easily measured than the store of data. Subtract those two from the Stats Canada estimate, and the value of data by itself ranges from $105 to $151 billion for the whole country.

Using share of GDP for our sector, I estimate that publicly-funded institutions such as hospitals and universities might be holding $6.6 to $9.5 billion in data assets, assuming all subsectors have been equally productive. (2)

Drill down farther, and my back-of-envelope calculation finds that Dalhousie University’s data might be valued between $24 and $35 million. (3)

There are too many layers of assumptions to feel confident in the accuracy of that range. (For example, it is unclear whether or where research data is counted in the national statistics.) Better than trying to decompose the national figures, we might try valuing our data using Stats Canada’s methods.

Our university’s data is not for sale, so a direct market value cannot be determined. Stats Canada values data at the cost of producing it, plus an estimated return on capital. Data is produced by people engaged in data-related activities: Working with many years of labour-market surveys, the agency identified roles likely involved in the creation of data (data entry clerks, researchers, analysts and so on), estimated a likely percentage range for the amount of their job spent on data creation, and calculated the cost based on salaries plus associated costs.

Value based on cost of creation sets only the lower bound; as new uses for data are found, its market value could change dramatically. More work remains to be done. But as a starting point it might hold promise for institutions seeking to measure their data as a financial asset.

In the system of national accounts, data may come to form a whole new asset class. For businesses, non-profits and other organizations, adding data investments to financial statements would give it added prominence.

Which brings me back to data governance. The value proposition for data governance at a university is hard to define. Counting data as an asset — literally — might be a step in the right direction.


  1. Statistics Canada has produced two highly readable papers on this topic: “Measuring investment in data, databases and data science: Conceptual framework,” released 24 June 2019, and “The value of data in Canada: Experimental estimates,” released 10 July 2019.
  2. Economic production in the non-profit sector totalled $169.2 billion in 2017, representing 8.5% of Canada’s gross domestic product (GDP). In recent years, publicly-funded institutions such as hospitals and universities have accounted for about 6.3% of GDP, which leads me to conclude that our subsector might be holding $6.6 to $9.5 billion in data assets.
  3. In 2017, education institutions in Canada accounted for $46.5 billion, or 2.3% of GDP, suggesting total data assets of $2.4 to $3.5 billion. Education institutions in Nova Scotia accounted for $1.4 billion, or 0.07%, of GDP – implying data assets of $74 to $106 million. Taking the size of Dalhousie University relative to all the others in the province, our stock of data could be worth between $24 and $35 million. (Thank you to Juuso Vesanto, BI Analyst with Dalhousie Advancement, for pointing me to an Economist report on this topic which gave me the idea to use share of GDP.)

Effective data governance requires a leadership mindset

The path to establishing effective institutional data governance might never be smooth, but some paths are smoother than others.

Success is more likely when people know their prime object is to advance the mission of the institution. Leaders recognize that data governance is the policy framework that establishes that key administrative data, regardless of where it is created or where it resides, is an asset of the institution, to be used to guide strategy and decision-making.

Success is less likely when people are primarily advocates for their area of custodianship or for a set of professional principles. Issues of confidentiality, security, data management, and protection of personal information are important, but do not define the core value proposition of data governance.

Advocates see policy through a lens that takes on the colour of their subject-matter expertise. Leaders appreciate the importance of these considerations, but strive to place those important pieces in proper relation to the value proposition.

The beauty is that advocacy vs. leadership isn’t about rank. You get to choose which role to play, and when. When you advise, you might do best to be an advocate. When you get to frame policy, you need to be a leader.

Data governance requires both mindsets, but ultimately leadership must prevail.

Two challenges for data governance in higher education

I recently listened in on an online meeting on data governance with about 140 others at universities all over North America. Two guest speakers presented two very different stories of the evolution of data governance at their institutions. It does seem there are multiple possible approaches.

I took away a couple of related insights about challenges unique to higher ed. Credit for these insights go to the presenters, San Cannon of the University of Rochester and Beth Prince-Bradbury of the Rochester Institute of Technology. Any unsupported elaborations or misinterpretations are my own.

One: The value proposition for data governance is not obvious.

Two: Cultural change is one of the biggest challenges.

Most experts in data governance are talking about the private sector. But higher education is not like the private sector. We are not profit-based. For us it’s not about ROI. We are mission-driven, and this makes it harder to define the value proposition.

Higher ed is also not a command-and-control world. Support from the top is not enough to overcome passive resistance. A Chief Data Officer who reports to the provost isn’t necessarily invested with a lot of power; he or she still faces the same cultural and change issues working with departments who consider institutional data “theirs.”

Challenges, yes – but these can be translated into useful guidance.

First, we must clearly define the value proposition of data governance in terms of driving the institution’s mission and strategic direction. Then we need to socialize that value as part of a robust change management effort.

One of the presenters noted that data governance is about policy but it’s not about enforcement. The model gains its authority and legitimacy not from presidential or provostial decree, but from the willing participation of people and departments across the university. Maybe the best model stresses leadership from the middle of the organization rather than the top. I don’t know.

What I do know is that a shared understanding of institutional mission and direction, desirable in itself, would smooth the way for effective data governance.

How do we know philanthropy helps students succeed?

I have little doubt that philanthropy supports student success. Just don’t ask me how many students receive meaningful supports thanks to donors, or how donor dollars correlate to increases in student retention. The way universities usually silo their data, those questions are nearly impossible to answer.

We ought to have answers. Reporting dollars raised is not the same as reporting impact. Increasingly, institutional strategic goals involve various departments working together. If we have shared goals, then we need to have shared data. How else to set targets and track progress?

We need Data Governance.

Data Governance is the policy framework that establishes that key administrative data, regardless of where it is created or where it resides, is an asset of the institution, to be used to guide strategy and decision-making.

Where governance is lacking, the problem is often perceived as primarily technical (give it to IT to figure out) or a negotiation among data “owners” (get all 50 stakeholders in the room). It’s neither.

IT and the individual custodians of data operate a level or two below governance. They deal with issues of data management, data definitions, security, protection of personal information, access control and appropriate use, and so on. All important matters, but secondary to what is fundamentally an issue of institutional policy, formulated at the level of university leadership.

So when I think of Data Governance, I assume it’s prefixed by the word “institutional.” The term doesn’t have meaning for me when applied to a single department such as Advancement. That might be data management, but it’s not data governance.

Given its central nature, what role can Advancement Operations play? To borrow a line from another context, we must lead from where we are. Where governance is lacking it is up to us to work with allies across the institution to give the issue the profile it deserves.