A recent report by Dimensional Research and Snowflake Computing suggested that as many as 88% of recent data initiatives faced failures. Earlier reports by McKinsey & Company and Gartner, amongst others, also highlight mixed results from data projects, with some spectacular successes and many expensive flops.
The Tessella paper draws on over 30 years’ experience of delivering business impact through data projects, consultation with industry giants and large organisations, and extensive research. Combining world leading expertise and hard-won experience, the paper recommends five clear steps to ensure data project success and to avoid common mistakes.
- Focus on business outcomes, not data
Successful analytics programmes start by identifying what the business is trying to achieve and what decisions must be taken to reach those goals. Only then do they assess what data and technology are needed to inform those decisions. We call this “Decisions First, Data Last”.
- Have a big vision but focus on quick wins
Many data projects fail because they are too big and take too long to deliver value, leading senior teams to lose interest. Data projects must have a pragmatic execution plan, with milestones designed to demonstrate early success. The first data project plans should focus on multiple, smaller projects, run with agility, to deliver the fast actionable results and rapid-fire value that will win over senior teams.
- Who, when and how will data be acted upon
Data success requires an understanding of who will use the data, when the information is needed and how they engage with the insights being provided. By doing so, the resulting insights are presented in an appropriate manner for the decision maker.
Project outputs need to be used by all sorts of people: it may be a data visualisation for an expert in drug chemistry or oil well drilling, or it may be a mobile app which presents complex analytics of multiple health metrics as a simple text recommendation. Getting it wrong may mean missed opportunities, lost customers and disillusioned staff.
- Replace silos with translators and collaboration
True business transformational data projects transcend traditional organisational boundaries. Companies need to adopt newly evolved structures, creating a culture where data scientists are in direct contact with the business functions, the IT departments and the communities to which they are providing insights.
Teams must be led by someone with a strong understanding of the both business context and technical challenges, these are the vital ‘translators’ who can speak the language of the business and data scientist.
- Take a scientific approach to data science
Many analytics strategies fail because they put technology first; Invest into an analytics platform, a black box, which may rapidly identify trends in their data sets. However, these correlations may not be meaningful in a business context. To deliver the effective insights, the reasons for these correlations need to be fully understood.
This is where a scientific approach comes in because it is about the identification, control and elimination of excess variability. Teams that understand the data and the industry issue being investigated, design and execute finely crafted experiments which eliminate variability and hidden biases. Thus dismissing accidental correlations, and reducing unnecessary averaging of the results. Through this approach, they can identify clear lines of causation between the decisions and outcomes, providing confidence and trust in the analytics and delivered insights.
Commenting on the paper, Matt Jones, Analytics Strategist at Tessella, says: “Good business decisions are those that are the best informed. Establishing causal links between data and the outcome allows you to move from intuition-based to evidence-based decision making. The right links can only be established by identifying the decisions that need to be made first and then setting up your data, technologies and teams in a way that ensures all work is constantly linked back to business outcomes.”
