From work, logistics and communications, to shopping, town planning and customer relations: our world relies on a constant flow of data. With vast amounts of data to call on, suppliers have never had a clearer view of their customers. This makes businesses more responsive and home life more entertaining, as customer profiling helps stores and streaming services more accurately recommend products and content to match our tastes.
It’s not the volume of data, but how you use it that matters
Amazon, which has become the world’s third largest company, at least partially on the back of its successful recommendation engine, showed that it’s not so much the amount of data an organisation holds that dictates success or failure. Rather, it’s how it chooses to analyse it. Instead of basing recommendations on the diverse, ever changing data set that is its current customer profiles, it uses product comparisons.
“Inspecting the recent-purchase histories of everyone who bought a given item requires far fewer lookups than identifying the customers who most resemble a given site visitor,” notes Larry Hardesty at Amazon Science. So, “even with early-2000s technology, it was computationally feasible to produce an updated list of related items for every product on the Amazon site on a daily basis.”
The rest is history. The human benefits of Amazon’s smart use of data drove its growth and, in 2020, it topped The Verge’s list of the most trusted tech companies.
Legacy systems and modern data requirements
Yet, while timely data will always be the most valuable, focusing on this to the exclusion of all else similarly has drawbacks. Reliance solely on data gathered during the last 18 months, for example, would cover only a narrow and unusual period, as we lived through the pandemic. Combining this with historical data would paint a clearer picture as we move into the new normal.
In many cases, historical data may be trapped in legacy systems, designed for stand-alone or mainframe use, not compatibility and sharing. This data must be rescued as, even if it has no immediate use, it can inform effective AI through more diverse machine learning. In many cases, this will require a bespoke ETL process, or APIs that will bridge the gap between incompatible systems.
If the value of data is to be fully realised, it must always remain accessible and portable. When it’s siloed or isolated, it cannot be used to generate revenue or for human good.
Data solutions for the greater good
This need for the easy flow of data is underpinned by the UK government’s determination to press ahead with its own global positioning and timing project since, post-Brexit, it’s no longer a member of the group developing the EU’s Galileo project, with its promise of diverse human benefits. “Such networks are seen by the government as critical for the functioning of transport systems, energy networks, mobile communications and national security and defense, while boosting the UK’s space industry and developing capabilities in these services,” reported Computer Weekly.
The UK government spent $500m on OneWeb and its fleet of 110 satellites, which would cover the UK and arctic region and support the country’s £300bn of economic activity generated by satellite services. It will do so by providing broadband connectivity and enabling digital services in areas currently lacking low latency broadband connectivity. It will also have applications in aviation, marine and emergency scenarios, as is the case with the rival StarLink constellation, which was used by the Louisiana government to aid recovery and rebuilding in the wake of 2021’s Hurricane Ida.
Data in education and further education
Such connectivity will be a boon for education too, with increased use of satellite broadband underpinning more robust connections for both schools and higher education. Satellite broadband is providing a vital connection for the 9m US school students lacking high-speed home internet, while Gilat is building satellite-based e-learning solutions for rural communities worldwide.
In an age where books and lessons, in pure data format, can be delivered through platform-agnostic LMS to desktops, tablets and mobile devices, connectivity is the key to unlocking universal education across all ages, including in-work and tertiary training.
However, we must remain aware of the fact that while teaching can be delivered using data alone, educational establishments are not mere learning factories. In many cases, they provide a level of pastoral support that data cannot replace unless supplemented by personal support.
Risks to the data-based economy
Intangibles, like pastoral care, aren’t the only area of concern. While predictions based on recent data facilitate not only increased sell-through, but faster turn-around and shorter supply lines, they can equally encourage a tendency towards just in time (JIT) supply or production. This reduces the capacity of organisations to mitigate for demand- or supply-side shocks.
“Because JIT production is based entirely on existing orders, it is not the most efficient system for dealing with the unexpected,” notes Investopedia. “This can mean extended delays, dissatisfied customers and potential forfeit of part or all of an order if any supply chain issues arise.”
AI and machine learning can reduce the impact, by identifying patterns in massive data sets to deliver predictive analytics. By forecasting shortfalls before they occur, they reduce the likelihood and severity of potential shocks – particularly when supplemented by third-party data to provide a richer, more granular picture.
Perhaps more seriously, while accurate profiling is useful in giving the business opportunities to increase revenues – equally benefiting clients as it presents services and products of greater relevance – it tends to reinforce existing behavior. The likelihood of making fortuitous discoveries is reduced and, if customers come to rely on a narrow range of products, there’s limited opportunity for suppliers to up-sell.
Each of these can be mitigated, and indeed it is essential that organisations do so – for their own sake, and the benefit of the humans with whom they interact.
Related Case Studies
Automotive Data Aggregation Using Cutting Edge Tech Tools
An award-winning automotive client whose product allows the valuation of vehicles anywhere in the world and tracks millions of price points and specification details across a large range of vehicles.
Formularies Data Aggregation Using Machine Learning
A leading provider of data, insight and intelligence across the UK healthcare community owns a range of brands that caters to the pharmaceutical sector and healthcare professionals in the UK.