This article first appeared in Digital Edge, The Edge Malaysia Weekly on October 11, 2021 - October 17, 2021
Decisions can be made in a few ways. It can be based on past experience, logical inferences, gut feeling or after one has analysed all relevant information.
The problem of relying on the first three options is that one can be misguided if the situation on the ground is not as one imagines it to be.
This is where data — defined as facts and statistics gathered for analysis — comes in handy. Of course, the way in which data is interpreted can be subjective. However, having accurate data can help business owners and policymakers strategise more effectively; it can also help the public make more informed decisions.
As more processes are digitalised, more data is being generated. In fact, the International Data Corporation (IDC) found that the amount of data created last year increased abruptly because more people were consuming entertainment and working from home.
Unfortunately, a lot of data is in a semi-structured or unstructured format, which means it is difficult to analyse using conventional data tools, according to various reports on the subject. Going forward, more unstructured data is bound to be generated from social media, videos, voice and sensor data.
The opportunity to leverage data insights and make better decisions, however, is there. Technology that can analyse unstructured data is also increasingly available. In this article, Digital Edge speaks to experts to understand how unstructured data can be managed in the private and public sectors — to make a difference.
During a workshop on data analytics with clients, Vernon Chua, co-founder and CEO of Innergia Labs, a retail data analytics provider, asked a participant who runs a food and beverage (F&B) chain how he thought his recent promotion campaign had gone.
The chain store had been running a campaign for customers to redeem an item on the menu. The participant thought the campaign had gone quite well based on what he had heard.
“But when we pulled up the data, it turned out that in a week, there were only five redemptions of that item. How could they have got it so wrong? Only when they were able to see the data insights could they understand what was going on and take corrective measures,” says Chua.
This is a common problem faced by businesses, especially those with more than one outlet. Many of them are surprised by what the data tells them, especially when it goes against their assumptions.
But it is becoming more important for businesses to manage their data properly and use those insights to scale. It is also seen as the next step for businesses after they have set up an online storefront, joined major e-commerce platforms or begun to sell items via social media.
The pandemic forced many businesses to digitalise their front-end systems. This did not just allow them to reach customers in a new way but also opened up the world of data insights to them.
Prior to this, much of their data, be it sales or inventory information, was not digitalised and most of it remained in paper form. A lot of it was considered unstructured data.
“Many small and medium enterprises (SMEs) start digitalising by adopting an electronic point of sales (POS) system. Then, they will plan for digital marketing and customer loyalty programmes. But if a business hasn’t even adopted e-POS, it is just sitting on a bunch of unstructured data,” says Reza Razali, CEO of Slurp!, a retail tech provider. “They don’t even know their customer data or what sells well. They only know how healthy the business is when the audit is done [at the end of the month].”
According to Khazanah Research Institute (KRI), most SMEs in Malaysia have digitalised their front-end systems by using computing devices and leveraging internet services but not their back-end processes such as inventory management, order fulfilment, cloud computing and data analytics.
SMEs risk being left behind in the digital economy, KRI observes, as larger firms can capture large volumes of data and implement their strategies more efficiently. This could result in large firms monopolising entire market segments because of their ability to extract, control and analyse data.
Consequently, SMEs should not think that having a digital presence and e-commerce storefront is enough.
“Having an e-commerce platform is not a silver bullet. It’s just another avenue for you to sell products. All you’ve done is change how people shop, but you still have the challenge of how to get them to come to your store,” says Chua.
Aggressive competition in the online sphere is the reason businesses cannot ignore data analytics, he adds. “You may not have the limitation of geography to reach customers now, but you have competition. They can buy from you or anyone else. So, what makes you stand out? There must be a lot of thought that goes into differentiating yourself from your competitors online.”
He gives the example of an F&B chain owner that used SYCARDA, his company’s data insights tool. The client analysed how its customers were paying and realised that it boiled down to credit cards from two major banks in Malaysia. The client took that data, went to the banks in question and negotiated a deal.
“One of the banks actually signed up and created a joint promotion on its credit card with the chain. It was so successful that, during the promotion, 40% of all transactions came from that one bank. There’s a lot you can do with data from just payment methods,” says Chua.
Another F&B chain used SYCARDA to monitor its best-selling items and realised that the 100 least popular items on its menu contributed 3% of its revenue but took up to 5% of its cost.
“In that case, what you can do is tighten your menu. Your chefs could focus on fewer items and when you open new outlets, it’s easier to scale and train new chefs,” says Chua.
When the pandemic hit, many of Reza’s new clients were just starting to move to e-POS systems. It was challenging. Without customer data, they only had a hunch about what was selling well in their outlets. If they wanted to open a digital storefront, should they sell all 100 items or just a few? These were questions they had to confront, says Reza.
Some of the SMEs were carrying out payments and sales via WhatsApp, Instagram or Facebook messaging, but this data was unstructured and could not be analysed.
“But even when SMEs have structured data, like their sales transactions, they don’t know what to do with it. All forms of data are underutilised by them at this point in time,” says Chua.
Regardless, getting an e-POS is the first step for a business to gain access to its data, and this can be integrated into digital marketing and inventory management insights later, Reza explains. Slurp!’s system provides information such as transaction data, products that sell well and average revenue per item to its clients.
“From our customer base, around 40% look at their data on a daily basis. Most of them just download it and give it to their accountants to be tabulated at the end of the month. It’s not fully exploited yet,” says Reza.
He recognises that SMEs might find it challenging to manage their data. But he holds out hope that with training, they will improve and start utilising their data to gain insights to strengthen their businesses.
“I notice that the focus of a lot of training programmes from the government has been on digital marketing. We should go beyond that now. If you do digital marketing on Facebook, you still need to rely on the social media company for the data. It’s important for SMEs to retain the data themselves,” says Reza.
Service providers like Slurp! and Innergia Labs could help SMEs achieve that. Slurp!’s POS system provides consumers with data insights whereas Innergia’s SYCARDA is a plug-and-play data analytics platform that costs less than RM150 per month for one outlet.
Other service providers can be found on the list of SME digitalisation grant service providers by Malaysia Digital Economy Corporation and the Selangor Information Technology & Digital Economy Corporation.
Ahmad Ashraf Ahmad Shaharudin, research associate at Khazanah Research Institute, often uses government data in his research. When there is limited data — for instance, granular or longer time series data — available, he will request more data from the relevant government agencies.
While the government, through the Department of Statistics Malaysia (DoSM), has published a lot of data in a machine-readable format, there is still room for improvement, says Ahmad Ashraf. Other governmental agencies collect and manage their own data, and much of this is even less accessible than that managed by DoSM, he adds.
“For example, agricultural data handled by the Ministry of Agriculture and Food Industries is mostly published in the annual Agrofood Statistics reports, which are available only in PDF format and not permanently accessible on the website.”
It is difficult for researchers to extract data for analysis using computers if it is published in PDF format. Machine-readable formats include comma-separated values (CSV) and Extensible Markup Language (XML).
This is important because open government data is crucial to maintaining transparency. Lack of transparency could result in misinformation and, in Ahmad Ashraf’s view, apathy in the community.
In fact, this matter gained attention recently when the new Minister of Health Khairy Jamaluddin promised to improve Covid-19 data reporting online, following criticism from various quarters. This data is used by the healthcare industry, government and civil society organisations to assess what needs to be done, while the public wants to know the current status of the pandemic.
In general, says Ahmad Ashraf, “the government is not sharing enough information. And when it does, it doesn’t share it in the most efficient and organised way”.
This could be due to a variety of factors. One is the lack of an overarching policy to compel government agencies to publish data in accessible formats. Another is that this activity is considered low priority and, in most agencies, only a small team is responsible for carrying out open data government initiatives. This was elaborated by Ahmad Ashraf in a discussion paper published in April.
These problems can be addressed by enacting a Right to Information law in Malaysia, he believes, which will serve as an overarching law on government data. The data must be published in machine-readable format and proactive data sharing on open data platforms must be encouraged.
In addition, privacy laws that extend to government data should be introduced to strengthen public trust towards the open government data agenda. He also suggests that the government streamline its data management system. “This could be done by empowering a designated agency to implement open government data with a cross-agency enforcement mandate. The government should also invest in data infrastructure and expertise across all agencies.”
Over the years, DoSM has been working hard to improve its data reporting capabilities, says Datuk Sri Dr Mohd Uzir Mahidin, chief statistician of Malaysia at DoSM.
Every report it releases now has to be in a machine-readable format, he says, and it has been trying to digitalise statistics uploaded in the past. It also creates infographics to accompany the data. All this can be downloaded at the Malaysia Informative Data Centre and from DoSM’s website.
“When we complete the census, we want to provide statistics for every member of parliament and state assemblies so they know how to represent their areas. We really believe that a strong nation requires strong data reporting. We want to empower people with statistics so they know what is going on and can speak up or participate in public policy,” says Mohd Uzir.
The department is also trying to upskill itself. “What was required of statisticians 30 years ago is very different from now. We need them to have advanced skillsets and be addressed as a data scientist instead of a statistician. They need to equip themselves with skills to not just look at data but also knowledge about different sectors. For instance, if you want to know about agricultural statistics, you must also know about agriculture,” says Mohd Uzir, who has been working in the department since the 1990s.
Asked about the suggestions posed by Ahmad Ashraf, Mohd Uzir acknowledges the pain, saying that DoSM is strengthening its system.
This includes revising the Statistics Act to strengthen the role of DoSM and streamline data coordination between government agencies. Strong data governance is important, he believes, as is the role of a chief statistician who can work across agencies.
The amendment of the Statistics Act was last mentioned in October 2020 by then minister in the Prime Minister’s Department (Economy) Datuk Seri Mustapa Mohamed. He also spoke of a National Statistics and Data Council that will serve as the highest agency in the country for national data and analysis, and a National Data Analysis Centre.
Mohd Uzir hopes the implementation of application programming interfaces and proper data management can overcome data-sharing barriers between agencies.
“Going forward, investments in transparency initiatives, data governance infrastructure and establishing a strong, digitally-trained workforce should be the top priority of any government agency and public sector organisation involved in the collection of data,” he says.
Save by subscribing to us for your print and/or digital copy.