The biggest challenge is improving the quality of supplier data © Getty Images
The biggest challenge is improving the quality of supplier data © Getty Images

Where to start when making sense of data

21 December 2020

We’re all familiar with the wide-reaching impacts of Covid-19 but these disruptions were exacerbated for businesses reliant on outdated technologies, as procurement teams fought to react to frequent, sometimes daily, changes. In response, many firms are investing in data analytics and intelligence tools to gain supply chain visibility and support real-time decision-making.

Supply Management spoke to data and supply chain experts at procurement consultancies GEP and Achilles to ask what are the foundations of effective data intelligence.

1. Start with clean, structured data 

Data-driven decisions depend on asking the right questions of the right suppliers, and collecting answers in the right way. It needs strong foundations of data governance and stewardship. 

Investing early in data governance pays by reducing the time spent cleaning data and limiting the likelihood of making the wrong decisions. Regardless, clean data always requires hard work. This involves good data structuring, understanding of the sources, collecting objectives, and the meaning of that data for the business. 

It is best achieved via good communication between the technical and the business teams. For example, avoiding the dreaded free text response is critical to achieving clean data. Wherever possible, the collection of data should be restricted to allowable response types that are either reference lists (like postcodes or counties) or fit a pattern (like an email address). This will also enable you to manage responses in multiple languages without having to translate twice. 

By collecting data in a structured, thoughtful way, you can avoid the time and frustration that comes with post-collection data cleansing.
- Achilles

2. Go into the supply chain

The biggest challenge for most companies is improving the quality of their supplier data. 

Today, most companies only collect data from their tier one suppliers. Instead, multinationals with multi-step, multi-geography supply chains need to capture critical data from suppliers and sub-suppliers two or three tiers up to truly map and manage their end-to-end supply chains. You can never have too much information to start with, so collect all you can, such as orders, inventory, shipments and supplier information, from tier two and tier three suppliers to put yourself in the best position to gain real-time visibility into the entire global supply chain.  

3. Predict consumer demand 

Information from the supply chain is only half the story. Customer demand drives everything. To inform future production schedules and purchases you need real-time consumer demand forecasting. Most organisations only project customer demand using historical sales patterns, which in 2020 were irrelevant by March. To overcome it you could conduct analysis of a combination of historic customer demand, purchase orders, internal sales goals and external third-party real-time data feeds. External data feeds, including public health information, weather forecasts, major events and social media influencers, can collectively provide a sense of consumers’ future demands.

4. Create valuable visualisations

To find the value in large, complex data sets, an organisation needs to understand what is valuable for the business and have the technical ability to throw complex questions at messy data. It requires individuals with a good mix of business domain expertise and advanced analytics skills working in tandem to achieve mutual goals. For data to be actionable, it needs to be easily consumed and understood. Data visualisation is critical to enable the decision-maker to query the data in order to answer their questions. The journey typically starts with identifying the key business problems and understanding what information could inform the solutions to those challenges. From there, data solutions should be auditable and repeatable, to ensure successful data solutions can scale up and remain a live reporting system. This will create business intelligence tools that track KPIs and enable key stakeholders to make data-driven decisions in a quick but rigorous fashion.
– Achilles

5. Establish control towers 

Bring all the financial plan data, customer demand-sensing data, and real-time supplier data together with third-party external feeds such as weather events and port shut-downs into one central data lake. Then use AI to scan for predictive patterns, stress test all nodes of the supply chain, and conduct “what-if?” planning in order to effectively mitigate risk.

6. Talent ensures future-proofing 

Flexibility and future-proofing depend on technology and talent. The choice of technology and the type of data/business are key in developing a future-proof environment. The baseline is to create an infrastructure and an architecture for data storage and data analysis that’s as flexible as possible, and to abstract and generalise ahead. Still, data intelligence tools are born, evolve and die every day – it requires human talent to rein them in, and the right attitude to continuously learn and adapt to stay ahead of the curve.
– Achilles

Fabrizio Margaroli, lead data scientist, Simone Gelli, senior data scientist, and Katie Tamblin, chief product officer are from Achilles. Allen Ozyazgan is VP of Supply Chain Solutions at GEP.

CIPS Knowledge
Find out more with CIPS Knowledge:
  • best practice insights
  • guidance
  • tools and templates