Weavestrands

5 actions for remarkable results from your data analytics assets

A recent IDC study projects a double-digit revenue growth through 2022 to $274.3 billion on Big Data and Business Analytics (BDA)  solutions. Other studies claimed that 60% of such projects failed. Even successful ones realised only 30% of the expected business value. 

This article discusses five actions that leaders have taken to get compelling insights and to ensure continued successful usage of analytics in their businesses.

1. Strengthen Data Governance to minimise the impact of incorrect data 

” The rotten apple spoils his companions. “

– Benjamin Franklin

The proportion of an organisation’s internal data (KPIs, Transactions, Master Data, Financials) to external data (All other accessible data in the world) has changed dramatically in favour of the latter. Indeed Jorn Lyseggen, in his book Outside Insight, argues that the worth of internal data is very minimal compared to the external data. Much of the external information is available in curated forms and public APIs. While internal data can be easily validated, external data could be compromised and manipulated.

e.g. A 2018 report from Blockchain Transparent Institute (BTI) on bitcoin trading found a high volume of fake transactions by algorithms and bots. Financial analysts who built analytics based on such data would have made erroneous inferences and disastrous decisions.

Given the increased sophistication of deep-fakes technology, it would be difficult to eliminate the usage of incorrect data. Organisations would need to strengthen and invest in data governance processes, and continuously scrutinise the gleaned insights for real-world validation.

2. Challenge Assumptions and data sources in the model 

The heart of a modern analytics system is the algorithmic data model. Advances in Machine Learning has enabled systems that can make assumptions, develop models iteratively while validating with available data and then provide the best possible insights to the decision-makers.

e.g. A FMCG company analysed region-wise demand and sales data for deodorants with a complex model including ambient temperature. The analysts augmented the model with data on humidity, from private weather stations, and transient weather conditions in coastal areas. The marketers got much more valuable insights to help tailor their campaigns through the changing seasons.

Organisations have to continually update the models for data that inevitably will be available more granularly (e.g. From single 24 hours total rainfall data to one in 1/2 hour blocks) and widely (e.g. One Rain gauge per town to one every square kilometre across the city). 

3. Ingrain Data-based decision making in the organisation culture

“In God, we trust. All others must bring data.”

– Attributed to W. Edwards Deming

Transactional / ERP systems are the lifeline of a business and are very visible in terms of processes (e.g. invoices, payments) and outcomes (e.g. total collections, sales/employee). It is impossible to operate at large volumes without such systems. In contrast, a hunch-based insight may happen to be the same as that provided by a sophisticated model with valuable data.

e.g. A Singapore based retailer looking to expand across the ASEAN region, chose to do so first in Indonesia and then in other countries. The data model suggested Indonesia as the first option and the team presented to the CEO, who was thrilled that it validated his instinct. 

Successful CEOs set the tone by continually questioning the hypotheses, data and insight. Gut, Instinct, and Hunches are essential in decision making. However, the sequence is critical. They should come up for discussion after the data analytics has made its point.

4. Ensure appropriate Usage of Data visualisation tools 

“You never get a second chance to make a first impression “

– Andrew Grant

Data visualisation tools like Tableau, Looker, Sisense are the medium of choice to present the outputs of Analytics systems. However, excessive usage of charts, colour and complex graphs can often detract the original insight. 

 

Tufts University study showed that people form a strong opinion on an infographic (data in a visual form with design elements) within 0.5 seconds of glancing at it. Another challenge is that a hand-phone, tablet or a laptop computer could be used to see the infographic for the first time and the impressions could vary significantly.  

 

Successful leaders use the power of the tools to look at the same data in multiple visualisations (e.g. Change the scale of a graph, Overlay multiple trend lines). Ultimately there is no substitute for looking at some of the raw data to validate the insight.

5. Skill and Re-skill for Analytics  

Kasey Panetta, in a Gartner article,  likens an organisation with no data literacy to a business where the marketing department speaks French, the product designers speak German, the analytics team speaks Spanish and, no one speaks a second language.

Data Analytics cannot be a separate department in a company. It is a continuous output of all the business activities and all available data. 

Skills to “source and model” data are essential. However, the skills needed to “consume and act” are vital to justify the investments. 

While organisations are increasingly investing for specialised analytics skills, they would need to do more to re-skill their marketers, manufacturers, buyers and accountants.

Conclusion

To maximise the return on their investments in Data Analytics, Businesses should  

1. Strengthen Data Governance to minimise the impact of incorrect data

2. Challenge Data sources and assumptions in the model

3. Ingrain Data based decision making in the organisational culture

4. Ensure appropriate usage of Data Visualisation tools

5. Skill and Re-skill for Analytics

Pic Credit : https://pixabay.com/photos/needle-in-a-haystack-needle-haystack-1752846/

 

   Send article as PDF