Startups and small businesses form a valued part of the European economy, and as such are supported by a wide range of incubator and accelerator programmes designed to help fledgeling companies on their journeys. Many common issues faced by startups are addressed widely, with help available in planning, HR, finance, marketing and fundraising – but the challenges facing data-focused businesses are unique.

Data Pitch, the first-of-its-kind EU-wide data accelerator programme, was created to respond to this problem, by bringing together large multinational companies that own data with startups and SMEs that can provide innovative solutions to common problems across specific industry sectors. In doing so, Data Pitch is creating a framework for better and easier co-operation, strengthening the EU data economy and creating economic growth.

However, there have been lessons to learn along the way. Below, we highlight 10 key issues we encountered while working with our 47 startup companies – and share the insights we gained about their needs, the blockers to successful data sharing and the issues they need help to overcome:

  • Build an interdisciplinary team. Most data-driven startups have a technical team, but don’t necessarily have people who have run a business before. The most successful startups we’ve seen have both types of people. Some startups address this by upskilling, learning the nuts and bolts of business management, while others bring someone in as a co-founder. A good example of this is Data Pitch startup Recog.nai, who brought together a team who had complementary skills (in tech, business, design) which allowed them to build a great product while smoothly running the business side.
  • Are you ready to use externally-sourced data? There are lots of things to consider when partnering with data providers. If the data being shared includes personal data, it will need to be secured. If you are not insured and there is a breach in your company, then you will have to pay a fine. You will need to think about the level of security you require for your data infrastructure. If you’re not sure how strong your infrastructure is, use our legal and privacy toolkit to find out. 
  • Take advantage of support from accelerators to build trust. If you want to work in machine-learning, you are relying on the accuracy of your algorithm. You can only do that if you have lots of data, which large companies tend to hold. There is often a trust barrier between the large organisation and the average startup, but accelerators can broker these relationships. Projects struggle to achieve their intended goals because of how long it takes to develop relationships with large corporates.


  • Be flexible with your planning. Be mindful of how long the process takes. If you need to agree on a one-on-one data sharing agreement you will need longer than you think, as well as the cash flow to support you. Accelerators like Data Pitch can help to move things along, helping advise startups on reducing the time spent on this process. Data Pitch created a data sharing agreement (from page 13 onwards) reduce the amount of time spent on the process.
  • Ensure you know who holds the intellectual property. Make sure you have a lawyer on hand, and ensure that you have a solid legal agreement in place. However lawyers can become expensive, so if you are early stage it is often better to join an accelerator who can provide you with legal counsel. Data Pitch created a contractual template (from page 16 onwards) to support startups in retaining the IP.


  • Relationship management. Make sure you understand each other’s role, as well as how much time the accelerator can dedicate to you. It is important to remember that they won’t be on call all the time. Get meeting dates in the diary at an early stage and make sure they don’t slip.
  • Read the fine print. Make sure to check for stipulations and constraints that may limit you in terms of future programme funding. Data Pitch startups could not, for example, apply for the European Data Incubator as this would have meant double funding from Horizon 2020. It is important to think strategically when applying for grant-funded opportunities, and find one that will add the most value to your startup.


  • It’s not all about the money. Though €100,000 equity free funding was an important incentive, the greatest value to our startups has been access to data. If you are involved in machine-learning, your biggest blocker is access to enough data. Funding can support many things, but data will bring the most benefits as it enables startups to deliver early stage proof of concepts. Data opens up a new world of opportunities that would ordinarily be off-limits to them practically and financially.


  • Ethical considerations. If you work in machine-learning, it is important to consider how  transparent you are. There is a spectrum of transparency, and while there is no right or wrong answer, Data Pitch has a privacy toolkit you can use to find the best way to proceed.

    Thinking about ethics means you are challenging unconscious bias, ensuring the inclusivity of your end product. In neural networking, the lack of human intervention means it is important to feed in enough data to remove the inherent bias. You need to consider the ethics of both the data being consumed and of the algorithm itself. The Open Data Institute’s Data Ethics Canvas helps identify and manage ethical issues, encouraging you to ask the important questions.

  • Domain expertise is crucial for startups. Even though lots of the startups had tech expertise, the most successful ones were those who had domain knowledge. They knew which problems they could solve for the companies, and in order to do so they required the data to be shared with them. By tackling a defined challenge, you are able to remove the uncertainty around market needs, and work with real data on real problems.