“When moving to a world of big data, development organizations — if they want to continue to be successful — have to embrace analytics and some of the opportunities they present.”
That’s what Justin Keeble, managing director of the sustainability services practice at Accenture Strategy, a global professional services company providing solutions in strategy, consulting, digital, technology, and operations, told Devex.
The term “big data” has quickly become a buzzword both within and outside of the development community and many such as the members of the International Alliance for Responsible Drinking as part of their Producer’s Commitments, are ready to embrace what it means to have access to analytics that will impact progress in global challenges such as tackling deaths caused by drinking and driving. But so many measures, programs, and technologies designed to interpret data can easily be daunting.
“Companies are using a lot of sophisticated analytical tools like Tableau, where they can do much more powerful manipulation of data and create better visuals to … understand the context,” said Keeble. “If you don’t do that, there’s a risk it can lead to misinterpretation or making bad decisions about what you think the data is telling you.”
With the increased pressure put on the value of data, how do we ensure the right outputs are being measured and the correct analysis is being done, while avoiding the pitfalls of excess data? Sitting down with Devex, Keeble explained.
This conversation has been edited for length and clarity.
How important is results’ analysis in ensuring development work is having the desired impact?
It’s absolutely crucial: If you think about the objectives of any development organization, they’re seeking to have an enduring impact and they need to design programs that will deliver on those objectives. So data is important in helping that organization deliver on its goals.
In 2012, the leading beer, wine, and spirits global producers made a collective commitment to build on their long-standing efforts to reduce harmful drinking through the beer, wine, and spirits Producers’ Commitments.
Each of the five commitments was measured and evaluated by a series of key performance indicators which were initially developed with advice from Accenture Strategy. The KPIs evolved over the five years to improve clarity and to reflect better the data acquisition process.
Data is being used to better inform where an organization’s focus is and to understand the nature of the problems they’re looking to solve. We’re seeing an increase in geographic information systems to understand how a natural environment or communities are being impacted to better target interventions, for example.
Data is important in making decisions about effectiveness, for example in delivering outputs versus inputs. Data can measure those inputs and outputs and make a comparison. If you’re training people around drinking and driving, you might run programs about being responsible when it comes to driving, but are you really having an impact on drinking and driving in that region or country?
What type of analysis yields the most valuable results?
There are three key things to take into account. First, there’s a challenge around consistency. Indicators and data need to be gathered and analyzed at the same time each year to ensure you can keep track of progress over time. I think this is another strength in our work with IARD members, as they’ve done a lot of work in gathering and submitting high-quality data each year. That creates a picture of change over time and our understanding of year-on-year progress. Ensuring you’ve got good quality data allows you to have consistency, which then allows you to have a picture of performance over time.
The second thing is about ease of understanding. In any analysis that puts progress into perspective, it’s a really powerful tool to create a valuable message to the audience — a statement that stakeholders can easily get their heads around irrespective of background and knowledge of the topic. You can have a lot of complexity in data analysis, but if you can’t understand what it means and articulate it in a simple way, you’ll fall short of helping stakeholders grasp and understand what the data is telling you.
“Ensuring you’ve got good quality data allows you to have consistency, which then allows you to have a picture of performance over time.”— Justin Keeble, managing director, Accenture Strategy, Sustainability Services
A third point is moving toward real-time insight. I see tremendous progress in technologies and the power of analytics. Real-time analysis will certainly apply to challenges development organizations are having in terms of what the data sets are that they use, what the insights they want are on a real-time basis, and how that data can be used to make more prescriptive and predictive judgements about what’s going to happen in the future. Development organizations will be able to shift from being reactive to proactive as they start to make judgements about the direction of travel in programs, initiatives, and the specific issues they’re trying to tackle.
What are some of the pitfalls to avoid?
First, I think you need to really test your metrics to understand whether the metrics developed are going to tell the right story and whether that’s what is needed. It’s very important to get feedback on metrics.
Secondly, this is a process that can take time. Make sure to give yourself plenty of time to gather the data, to ensure you can build in the right checkpoints to check accuracy and quality of data so you can go back and check anomalies. This is often a common source of delay, so building in that time for data cleansing is going to be important.
Also, make the process of data submissions as easy as possible. We really need to move toward a world where data collection and data entry is easy. What we’ve done over the five years with IARD members is identify processes and templates. Feedback we’ve had from signatories is that they’ve really appreciated the improvement of that user experience in populating the fields. We’ve done some work on prepopulating some of the data where we’ve got it available to try and make it much easier for the organizations to provide that data.
Fourth, I think for organizations that generate a large amount of data, the first pitfall is to have all this data but not know what they can do with it. We would strongly advocate starting with the problem you want to solve. What are the problems the development organizations are looking to solve? Frame that as the question, so you can be very clear about the problem you want to interrogate data to solve. If you do that the data works for you and you can be clear about outcomes — if not, you can spend a lot of time doing analysis and not necessarily solving the problem.
Finally, think about the value that this will deliver for the organization. If you can be clear on those outcomes that can then support the actions you can take.