Data, the cornerstone for successful digital transformation in R&D
Digital technologies are evolving so quickly it's hard to keep up. It seems there are almost daily breakthroughs, with highly publicized stories of artificial intelligence (AI) doing everything from predicting cancer growth to mimicking the human brain. New applications are constantly being found in industries like healthcare, finance, transportation and many others. It is not surprising that with all of this burgeoning potential, R&D organizations are also eager to jump in, making major investments in this space. However, there is a problem. The success rate on these projects is estimated to be as low as 15 percent! A recent example is the highly publicized setbacks in the application of IBM's Watson technology to healthcare.
So, what is the answer? If you wait, you risk falling behind competitors who achieve advantages with digital technologies. How can you make smart investments that position your organization for success with today's technologies, such as deep learning, as well as the next breakthroughs? This blog will show you how building a solid data foundation is essential to long-term success in implementing digital technologies and how you can get started with smart investments today that will pay dividends long into the future.
Garbage in, garbage out: How disorganized data could cost you
One reason for the high failure rate of corporate digital technology projects is that organizations become enamored with the hype surrounding the possibilities offered by new technology like AI—and overlook existing gaps in the less exciting foundational elements that are required to make these technologies successful. This oversight, coupled with overly ambitious expectations for AI's capabilities, can often lead to disappointing results.
One of the most prevalent gaps that organizations overlook is the strength of their underlying data assets. Ensuring that you have a high-quality, robust data foundation on which to build and train these digital technologies is critical. The prevalence of this "data blind spot" is evidenced by a recent survey by Forrester, in which 83 percent of respondents said that having a well-curated collection of data wasn't their biggest concern.
So why is high-quality data so important to success in the realm of digital technologies? The key is that these technologies cannot create any new knowledge. They learn and develop based on information that you provide. All of the outputs are linked to the quality of the inputs, i.e., your data. Thus, if you input incomplete, unstructured or irrelevant data, the quality of what is returned will suffer greatly, and if relied upon heavily, could create significant organizational risks for success and even safety. In fact, poor-quality data is estimated to cost the U.S. economy $3.1 trillion annually from misguided decisions, inefficiency, lost revenue and more.
Given this, how do you build a successful digital transformation strategy for your organization? Download our latest whitepaper.
The building blocks of a high-quality data foundation
Building a high-quality data foundation for your organization is the most critical investment to make when initiating a digital transformation. But what is good enough? The definition of what makes data high quality is broad, but we've identified three main attributes against which to measure your data readiness:
- You must have enough data. Digital technologies are hungry for as much data as you can give. Only with a large dataset can they come to accurate conclusions. For example, facial recognition on Facebook has an accuracy rate of 97.35 percent, which closely approaches human performance. This was only made possible with access to a dataset of 4 million Facebook users. The amount of data your sci-tech R&D company needs depends on the nature and complexity of the problems you are trying to tackle. Consider the needs of your intended future applications when defining your data requirements.
- You must have diverse data. Having a broad and deep data collection related to your organization's core competencies and markets is critical. However, if you intend to use digital technologies to look for growth and innovation opportunities, diversity of data is also very important. If your data is only related to an area that you already know, technologies such as network analysis or AI are not likely to uncover many new opportunities but will simply validate your current knowledge. Thus, it is important to identify and include data collections that are adjacent to your core focus.
- You must have well-organized data. To be useful for drawing actionable insights, it is crucial that your data collection is well structured. Building an overarching data model, curation and clean up is especially important for uniquely complex scientific data that may include text content, chemical structures, drug/target relationships, taxonomic animal names (kingdom, phylum, genus, etc.), financial values, schematics, graphs, charts, etc. Though statistical analyses and computational algorithms can be useful in organizing and enriching your large datasets, it is important that you also allow for human curation within your data governance process. Experienced chemists, biochemists and data scientists can analyze your data and offer insights that no AI system can.
Rapidly ramping up your data foundation
Although developing your data may feel like it's slowing down your progress, it is an investment that is vital to the success of any digital technology project. With technology evolving so quickly, a solidly built data foundation will give you the freedom to more efficiently implement each new AI and machine learning application, providing an advantage that will continue to provide returns for many years to come.
Under a tight timeline or facing talent gaps that are impeding your project? Consider partnering with external experts who can provide high-quality, well-structured data collections to build your data foundation faster and expertise to help you build the necessary structures and processes efficiently. At CAS, we have been managing, curating and extracting insight from scientific data for more than 100 years. Ready to get started? CAS can help.
Download our whitepaper on Profitable Digital Transformation in R&D to find out more or contact us to discuss your specific needs.
CAS, a division of the American Chemical Society, partners with R&D organizations globally to provide actionable scientific insights that help them plan, innovate, protect their innovations, and predict how new markets and opportunities will evolve. Leverage our unparalleled content, specialized technology, and unmatched human expertise to customize solutions that will give your organization an information advantage.