Data science is king
Artificial intelligence and machine learning in bid management is an emerging area of interest and its application is changing virtually every industry and business function. Not since the industrial revolution have we seen as much advancement in such a short window of time. Data science is now king, and it will be the driver of procurement 4.0.
With technology adoption and innovation being accelerated, every organisation is becoming a technology business. To keep up, procurement is fast tracking digital transformation and looking at new ways to be more future-oriented and shock resistant.
What will happen to the competitive bidding environment and what are the real business opportunities that artificial intelligence and machine learning technology can offer suppliers to build their competitive advantage?
“We cannot solve our problems with the same thinking we used when we created them” ~ Albert Einstein
Procurement has typically been slow to adapt to change without a trigger event, and Covid-19 was the critical inflection point that highlighted weaknesses and the consequences of relying on outdated legacy systems, antiquated procedures and fragmented sourcing processes.
As cyclical patterns show, when procurement wields its power to force system change – examples being paperless e-procurement for sourcing and the incorporation of sustainable development goals in decision criteria – suppliers follow. While there has been a lot of focus over the last few decades on procurement spend and cost savings, there has been little digital advancement in the way buyers and suppliers operate behind the scenes.
Now, accelerated changes are taking place to shift procurement from tactical sourcing to one of supply chain and social value creation enabled by automation of previously manual tasks, real time insights and analytics, and artificial intelligence that uses data models to streamline operations and support decision making.
Structural procurement changes and the impact on global bid management
Rapid globalisation, corporate governance and regulations to make public spending more transparent is making procurement tenders more frequent and complex in most markets.
As procurement departments restructure into centralised functions responsible for category and sub-category sourcing and spend management (such as ICT > Cloud Services / Hardware; and Professional Services > Legal / Engineering Services), more pressure is being placed on suppliers to compete for service bundle contracts.
This is especially challenging for organisations that need to respond to centralised procurement opportunities but have corporate structures that span business lines, divisions or countries, and even more problematic for organisations that have multiple entities and/or which have acquired new businesses.
How information or knowledge is stored, exchanged, traced, and cross-sold in the bidding process can have a major impact on both the speed with which an organisation can adapt to market competition and structural changes; and the relative success they will have in achieving optimum bid process efficiency, compliance, transparency and commercial impact.
Stages of digital maturity
“The future is already here – it’s just not evenly distributed” – William Gibson, author and originator of the term ‘cyberspace’
While there is no shortage of familiar frustrations and global shared experiences surrounding procurement, consultants and analysts such as Accenture, Deloitte and Gartner – among many others – predict that we are not too far away from a digital procurement model similar to Amazon that provides a simple and intuitive buying experience, and where data underpins and provides contextual information for both buyer and supplier relating to the process steps they need to take across each stage of the bid and contract lifecycle.
As procurement entities move towards ‘digital first’ centralised sourcing approaches, digital marketplaces that deliver greater visibility into category and supplier segment spend are a practical step-change, however reaching digital maturity and the ultimate Amazon model is neither a predictable destination, nor a consistent pathway for all. This is especially the case for the public sector, which, as a manager of taxpayer funds, has high expectations to deliver public value while being accountable for improving community outcomes. This is not an easy algorithmic problem to solve given the many bidding variables – such as company and project characteristics, competition, contract and regulatory conditions that impact both bidding strategy and supplier selection.
Use of data as a strategic tool however, can help to improve the efficiency and competitiveness of organisations and is particularly suitable for both suppliers and public procurement agencies where open data is available. And while bidding organisations that participate in competitive procurement often define the success of their efforts by their win ratio, little attention has been placed on how bidding inputs, decisions and outcomes data can be more optimally used for research, qualification and selection of opportunities that have the highest probability of success for both sides.
Applying digital pivots to increase digital maturity
Depending on procurement’s motivation and where it sits along the digital maturity curve will influence how it approaches the market, and how it expects the supply market to respond. Digital transformation doesn’t happen instantaneously – or equally – for all. There are, however, already signs that the procurement industry is preparing their workforces for change, with early adopters aiming for the north star – procurement 4.0 – where digitisation and advanced technologies and to some extent artificial intelligence and machine learning will enable full integration of a supply chain ecosystem where participants can model scenarios and adjust how they respond in real time as conditions change.
According to Gartner, this optimum state is called hyperautomation, which “deals with the application of advanced technologies, including artificial intelligence (AI) and machine learning (ML), to increasingly automate processes and augment humans.”
While no single tool can replace humans, the benefit of hyperautomation lies in how it can be applied to use cases like procurement 4.0 where discovery, analysis, design, automation, measurement, monitoring and reassessment is required. Getting there begins with small incremental steps and human buy-in to begin the transformation journey, and for bidding organisations, it starts with silo-busting, adapting to new digital standards, and then building in mechanisms to collect, trace and use the data.
Laying the foundation
Procurement opportunities are opening up to become easier to find and participate in – leaving the ingrained print culture and outdated content production process behind. In the UK, digital accessibility regulations apply to public sector web content whether it is open to the public or for internal use, as well as systems and content purchased by them.
For this reason PDF and multiple file upload and downloads will become less standard (or acceptable), and browser-friendly HTML formats that can be read across multiple devices will take over by default. The format of HTML allows for easier maintenance, improved accessibility and usability across devices, and compliance with open standards allows anyone with an up-to-date web browser to view the content without needing additional software.
HTML also captures information and analytics about how people are interacting and viewing content, and what links have been followed or shared. Additionally, HTML can be easily searched, translated, updated, printed and when required, and autosaved to accessible PDF.
Moving from guessing to knowing with data insights
Adopting a more data-driven culture and harnessing the power of data for business intelligence can be a game changer for organisations looking to begin or speed up their digital transformation journey. The more that data can be analysed, modelled and incorporated into day-to-day workflows, the more rapidly organisations can innovate and make decisions on the fly without relying on gut feel.
Collecting and analysing data in bidding and procurement can reveal trends and metrics that can help to increase the overall efficiency of a business or system that would otherwise be lost in a mass of information. Examples include identifying the proportion of a bid that has utilised content library over tailored content; repeated bottlenecks in the bid delivery schedule; or the average time contracts take to go from procurement planning to tender, to contract award and implementation.
Getting people used to working with structured data is a good first step in the digitisation process as it lends itself to an array of common use cases where querying and reporting can inform planning, scheduling and management of bids. Analytics can also help to augment decision-making by generating insights and recommended actions.
Enhancing speed, precision and effectiveness of human effort
Natural Language Processing (NLP) and machine learning are subsets of AI – the umbrella term that can mimic human intelligence to understand data, find patterns, make predictions and find recommended actions without explicit human instructions. Use of AI tools in bidding has to date focused mainly on time saving tasks that improve search and matching of content, and reducing the manual effort involved in document planning and set up. Advances in machine learning, however, now deliver the ability to build a system that can close the communication gap by learning from past experience and identify patterns that are not always immediately oblivious to a human.
For AI tools to be effective in a business context, they require some human supervision, or ‘supervised learning’, and a high degree of data science requirements to build the deep learning model. This involves actively training machines to perform a specific task, such as observing large data sets that have been classified into different categories. With a clear goal and some training data to begin, the AI can learn the algorithm and improve its accuracy (or confidence) over time to classify and categorise new data without human input. Examples of applications include:
Descriptive and diagnostic analytics
Descriptive analytics can be used to help executives draw conclusions on why certain business changes have taken place, such as correlations of bidding activity during holiday seasons or following public policy announcements, while diagnostic analytics can help to understand why events have taken place, for example data mining, correlation analysis and drilling down can provide reasons for bid losses, or an unusually low number of no bid decisions by a bidding company or bids received by the buyer (was the timeframe too tight to attract quality bidders and compliant submissions).
Predictive analytics, which uses statistics and machine learning modelling to predict future behaviour, is another application that builds upon what it has already learned from hindsight by identifying when patterns are likely to reoccur. For example, if a buyer is notorious for publishing procurement notices during holiday seasons, or has a tendency to issue dozens of addendum, predictive analytics can be used to help bidding companies identify and prevent potential risks, and help them to take advantage of future opportunities with proactive scheduling and human resourcing strategies.
Prescriptive analytics uses machine learning to analyse raw data in aggregate to help organisations make better decisions and a course of action in a situation. Factoring in possible scenarios such as available bidding resources, past win/loss outcomes, past project and/or KPI results, and current project pipeline can help organisations uncover insights that conventional analysis could not otherwise detect. Data modelling can also replace cognitive bias and flawed assumptions with fact-based insights about a project’s statistical chances of success. By analysing team composition, project size, and contract value, analytics can assess probabilities of project outcomes. This can help to improve the bid/no bid qualification criteria and assessment of the attractiveness of a given project, reassign resources or rebalance business portfolios away from underperforming projects.
Bottom line business benefits
According to McKinsey (2019) beyond automation, innovative digital solutions on the procurement side alone can unlock as much as an incremental 3 to 10 percent in annual cost savings. However the benefits run deeper. OECD (2017) research found that up to 50 percent of a contract value can be lost due to mismanagement and corruption, and with $15 trillion a year being spent on goods, services and works this is not insignificant.
By opening up data and making it transparent and traceable, other benefits include improving business continuity to match supply with demand, identifying critical suppliers in distress, and detecting purchasing patterns and red flags in behaviour, such as unusually short bid windows or regular contract price variations.
Conversely, bidding organisations can utilise data to optimise bid processes and support corporate strategy using data discovery and mining, correlation analysis and drill down to diagnose why certain events took place (was the win to loss ratio low because of poor bid qualification, misaligned solution or pricing strategy; or is it symptomatic of something greater within a business line or team).
Digital transformation cannot be a siloed effort. As organisations rethink their bidding strategies to adapt to future procurement go-to-market strategies, new generations of software that offer deep, specialised solutions will emerge which can easily integrate into the broader supply-chain ecosystem. It may take some time to get to the ultimate procurement 4.0 because in the end, it is humans who will decide how fast we are to go. And it won’t be the automation of workflows, the digitisation of paper and network connections, or even the business intelligence derived from the system (whether that’s through data science, artificial intelligence or machine learning in bid management) that will return the greatest value and competitive advantage – it’s what and how humans use it that will matter most.