Of all the sectors I’ve worked in – automotive, aerospace, retail, construction and housing – the last two face the greatest barriers when it comes to smart procurement technology.
Pressured by tiny margins and led by giants such as Amazon, the retail industry has perfected the art of getting the right products in front of the right customers, making the purchasing process as frictionless as possible thanks to chat bots and machine learning.
The automotive sector has used digital systems, namely enterprise resource planning (ERP), to integrate tightly with its supply chain. Car makers provide advance notice of manufacturing schedules to suppliers almost one year ahead, offering more granular detail as delivery dates approach.
These levels of efficiency in the retail and automotive industries are propelled by a critical need to reduce supply chain costs and never lose a sale. But the housing and construction sectors don’t have the same drivers. Margins aren’t quite so tight. Nor is the difference between profit and loss. And the supply chain complexity that drove digital procurement in aerospace and automotive isn’t mirrored in the more linear supplier ecosystem of the building world.
Regulation is another factor. The asset management and building sectors are incredibly tightly controlled, for good reason. The procurement mistakes around Grenfell illustrate how vital buying regulations are and also the high degree of perceived risk and cost attached to changing any tendering, sourcing or transactional processes. There is a real sense of the personal cost of failure that might come from change.
As a result, housing and construction organisations set targets that support the status quo and there is a culture of risk being transferred to supply chains. These factors all contribute to an innovation deficit in the sector, and the uptake of digital procurement technologies has suffered.
Another cause is poor data. Building and housing haven’t been as strong as other sectors in terms of gathering and categorising data and this has delayed smart analytics taking off. Up to recently, very little granular detail was collected, so it’s hard to introduce machine learning for predictive analysis or to compare spend data with other datasets to see if an organisation is paying too much.
Another barrier to adoption has been size. For data-led procurement technologies to be effective, hordes of analysts are needed to interpret data and draw meaningful insights. Procurement functions in the car manufacturing, retail and technology sectors have the budgets to employ these data scientists but that’s just not the case in building or housing.
Less expensive, boxed procurement technologies might just be the answer. Vast amounts of data can be thrown into analytical tools such as IBM’s AI system, Watson, side-stepping the need for large data analysis teams. Plug-and-play integration hubs like Jitterbit are flexible and easy to use, allowing businesses to join together with their supply chain’s technologies at low cost.
Machine learning is also on the rise in construction and housing procurement. As the sectors gather more refined, low-level spend data, algorithms will be used to classify this data, spot patterns and predict future trends.
For example, in my organisation, PfH, we can now identify how much a housing association is spending on, for instance, reinforced plastic bath panels; whether they are paying more than last year and if they are buying the best value panels. This granular data is matched against external data sets to see if that housing association is paying more than their peers.
Business analytics services such as Microsoft PowerBI are another increasingly popular, relatively low-cost smart technology being used by housing and construction organisations to connect different data sources. For instance, customer relationship management records about sales, renewals and contracts can be ‘plugged’ into live transactional datasets. This automatic bridging of statistics is reducing error, particularly around spend forecasting, giving organisations a multi-perspective view.
But this is just the start. By 2025 I predict that chatbots, AI and blockchain will be working together to fulfil significant chunks of the buying process in housing and construction. Big data and AI will provide insight into all purchasing activity, informing stocking habits; weighing up global, macro and micro economic activity to determine best time to buy; feeding into production forecasting and product criticality.
The question is whether humans will still have a part to pay in the procurement process of the future. In housing and construction, once supply chain and block chain relationships are established, there will be less need for human intervention, particularly around the administrative side of transactional procurement. But there will always be a requirement for buyers to manage the more strategic, nuanced work involving complex judgement calls and delicate relationship-building. That’s something smart procurement technologies haven’t yet replaced.
Phil Moss is chief technology officer at Procurement for Housing