How to make a business case for open data
Many organisations are interested in opening their data but are unsure of the associated costs and benefits. Here we guide you through the steps needed to maximise the benefits of opening data by asking four questions:
- What are your organisation's goals?
- How is data used now, and how could it be used?
- How can you get value for your business?
- How do the business models compare?
What are your organisation's goals?
The first step in working on a business case for open data is to identify your organisation's goals and the way in which they are satisfied. Open data should be used as a tool to help your organisation achieve these goals. These goals might include wider social, environmental and economic benefits as well as direct benefits to your organisation.
You should also consider how your organisation will work in the future, when much data is shared as open data. What role will your organisation play in that environment? Will there be new opportunities that you can take advantage of?
How is data used now, and how could it be used?
The second step is to look at the way data is used in areas where sharing data more easily would help your organisation meet its goals. Identify:
- the groups with whom you are sharing data
- data that is currently being shared between those agents
- problems with the currently shared data that makes it less useful than it could be
- other data that, if available, would help address your organisational goals
Potential groups involved in your data flows might be:
- other parts of your own organisation
Look at your existing data sharing arrangements and the numbers of requests for data that you receive to identify places where there is a demand for open data. Ask those groups whose work complements yours how they currently use data and what other data would enable them to work more effectively.
Organisations often underestimate the information they hold and the potential of that data.
- The main type of data that organisations work with is administrative data. These are the records that are made as part of an organisation carrying out its day-to-day business. Examples are point-of-sale receipts, website access logs, or railway maintenance plans.
- Organisations often own reference data that is simply used to provide information that rarely changes about things, which helps to understand other data. Reference data commonly includes identifiers for these things. Common reference data is often shared between and used by many organisations. Examples are product information, charity registrations, or taxi licences.
- To make sense of the low-level information within administrative datasets, they usually have to be analysed and summarised. This aggregate data can show trends over time or highlight differences in different geographical areas or for different products or services. An example would be total sales of each product over time across supermarkets.
Sharing raw administrative data can be hard because it often contains personal data and therefore can't usually be released under data protection legislation. Having unique access to this data may also provide you with a competitive advantage. There are usually fewer barriers to providing reference and aggregate data as open data.
It is easiest to share data that your organisation has gathered and maintains itself, because then you know that your organisation owns that data. If some of the data originates from another organisation, your ability to share that data will be determined by the licence through which it was made available to you. Open data can be shared (and if it was made available under a share-alike licence, you must use the same open licence for the data you have created). Closed data usually cannot be shared onwards.
How can you get value for your business?
When you look at what data is flowing between your organisation and other groups, there will be some that your organisation can affect, by sharing more or better quality data with others or by ensuring that you receive more or better quality data from someone else.
Publishing open data can be supported through three general business models, described in detail below:
- freemium: you provide an "added value" data product or service, for which you charge
- cross subsidy: you reach more customers, or provide enhanced services to existing customers, through wider sharing and use of your data
- network effects: by collaborating with other organisations, you reduce your costs in maintaining data which you use in your work or extend the possible audience for your products and services
These different ways of supporting open data can be used together for a single dataset, or different models can be used for different datasets. There isn't one right model to use. For example, an organisation might provide open access to some data to enable a larger community to improve its quality, exploiting network effects, and because it drives business towards the products and services offered by the organisation that owns it, which is a cross-subsidy-based business model.
Freemium business models are used for a variety of web applications. With a freemium business model, the free product or service is subsidised through a paid-for product or service that offers some kind of added value on top of what's made available as open data. The free product acts as marketing, establishing the provider in the marketplace and increasing the take-up of the paid-for product.
One way of using a freemium model is to release your open data using a share-alike licence. This ensures that organisations who do things with your data have to either openly share their results (which means you can benefit from what they do) or have to negotiate with you to be able to use the data under a different (potentially charged) licence.
OpenCorporates uses this business model, licensing their database with a share-alike licence while offering paid-for licences for companies who do not want to share their data.
Another approach to a freemium model is to offer a paid-for product that:
- incorporates additional data, perhaps from third-party sources
- is provided in a different format from the open data
- is more up-to-date, complete or detailed than the open data
- is the result of an analysis or model based on the released open data
- is a dump of data that can otherwise be accessed through an API
Alternatively, you could offer a paid-for service based on the open data you are publishing that:
- provides an API over open data that can otherwise be accessed as a dump
- provides availability guarantees through a Service-Level Agreement
- removes rate limits
Cross-subsidising business models fund the release of open data through other benefits to your organisation. Sharing open data can:
- help or prompt customers to use the products and services that you offer, which can help you gain and retain customers
- enable information to flow more efficiently within your organisation and between you and your partners, which can save time and resources within your organisation
- increase brand awareness and enhance your reputation
Here are some examples:
- advertise: datasets that are widely used can be effective advertising for your organisation, either due to attribution or by driving people to you using the identifiers in the data
- provide support services: as you are the people who know your data best, you could offer consultancy or other support services that help other people take advantage of the open data you provide
- charge for changes: if your dataset holds information from other organisations or individuals, and they benefit from having that information be up to date, you can charge them for changes they make to the dataset
- be accountable: openly sharing data about your organisation shows that you value transparency and accountability for your business practice, which can enhance your reputation
- reduce your data sharing overhead: if you have to regularly send the same information to multiple recipients, publishing that data can help them access it when they need it and reduce your internal administration burden
Making data available as open data lowers the barriers for third-party developers to reuse that data. Third-party developers can then create tools that:
- extend your audience: if your organisation has a goal of raising awareness of a particular issue, opening up data is one way to achieve this, particularly as third-party developers might provide access to the information through different channels (such as mobile apps)
- smooth demand curves: if you provide a service, sharing information about how busy you are at different times of day can enable third-party developers to create tools that help customers contact you when you are less busy, smoothing out demand for the service
- become more findable: if you have branches from which you provide a service, sharing information about their location, opening times, and what they offer can enable third-party developers to incorporate this information into maps, helping customers to find you
- out-source your R&D: opening up data can be a cost-effective mechanism for out-sourcing research and development of new products and services, which you can then take advantage of or purchase
The business models that have the most potential for a big return on investment, for all concerned, are those that take advantage of network effects around data sharing.
There are two kinds of network effects that you should try to take advantage of in an open data business model:
- Groups of organisations can collaboratively maintain open datasets, benefiting from each others' contributions and improving the set of data that each organisation can use. This can work particularly well to improve the quality of datasets that aren't complete or completely accurate. Contributions might alternatively include sponsorship or donations that help to support the dataset's maintenance.
- Releasing open data can help to grow the market for particular products and services, particularly when particular tools or services are well-placed to help organisations and individuals to take advantage of that data.
How do the business models compare?
You might identify several potential business models that involve publishing open data. To make a business case, you need to compare these with each other, with the impact of doing nothing, and (sometimes) with a business model based on selling the data.
Morten Lind's presentation on Value of Information: Assessing Social and Economic Benefits in the Context of SDI and PSI shows this kind of analyses in the context of a public sector organisation releasing various geographic datasets.
Organisations can overestimate the return on investment of selling the data they own (including making it available through a freemium business model) if they underestimate the costs that can be involved, such as:
- legal costs of creating and enforcing restrictive licences
- development costs of restricting access and use of data
- administrative costs of issuing licences
- sales and marketing costs to promote the data
There are also strategic risks if you build a business based on charging for data. As we have seen in the music, news and software industries, people are becoming less prepared to pay for digital products that can easily be copied and shared with others. Competitors or communities might also be able to undermine your data business by releasing their data as open data.
You might decide to trial using open data to discover what impact it has. If you do this, you should collect evidence during the trial to measure how much the open data is used and how this usage relates to value for your organisation. Bear in mind third parties might not invest in using data from your organisation unless some guarantees are given about its continuing availability and consistency over time.
Further discussion of business models for open data can be found at:
- Open data: Where's the Benefit? talk by Jeni Tennison
- Business model innovation open data slides by Harry Verwayen
- Business models for PSI re-use: a multidimensional framework paper by Enrico Ferro & Michele Osella