A value gap in analytics
Embedding analytics isn't about measuring employee productivity, it's a business-wide statement of intent
20 December 2021 | 0
In association with Aon
Embedding analytics helps people move naturally from insight to real-world action. Think of a busy colleague in operations that is expected to make decisions in real time by consulting analytics dashboards. If that dashboard is, say, three of four clicks away in a system behind yet another login screen, the motivation to take those extra steps will wane as they naturally default to their existing knowledge for a speedier decision. Whilst analytics strategies have planted deep roots across many industries, there remains this ‘value gap’ when our analytics are treated as offshoots of core business operations.
In 2020, Forbes reported that “only 13% of finance organisations are using artificial intelligence, analytics, and automation to transform multiple processes across their enterprises”. To truly drive value, analytics must therefore transform the process, not sit adjacent to it. This paints a bleak picture of our scenario where dashboards get lost in the hands of busy people with urgent decisions to make, never quite reaching the potential that their data promises; another bookmark in someone’s browser.
To change this, analytics must drive action back into the core value chain of a business. This means embedding analytical insights in new forms into the existing system and process landscape of the firm. This new challenge brings into focus the task ahead for firms that are accelerating their digital transformation and is sure to test us all in new and interesting ways.
Aon is in the business of better decisions, helping our clients in four key areas:
• Navigating new forms of volatility
• Building a resilient workforce
• Rethinking access to capital
• Addressing the underserved
Our data-driven solutions in these areas can come in many forms but share a common goal. We are committed to understanding the best ways to deploy analytics that naturally permeate and enrich the engagement between our colleagues and their clients. This begins with understanding the business context of the solution, whether it’s providing intelligence in distribution (sales & marketing) processes, in business transaction services, in client-facing advisory etc.
Only with an appreciation of this context can we begin to design solutions and define our options for augmenting the intelligence in decision-making, such as:
- Recommendation engines. Driving flexible decision-making within business systems by swapping locally coded logic with a probabilistic model, presenting a user with a recommendation among various choices.
- Unstructured data cleaning. Extracting key data from documents, linking it to the firm’s reference data and presenting the proposed records for validation to colleagues before releasing them deeper into the organisation.
- Action-oriented diagnostics. Visualising events or a position a client holds relative to some broader universe of data (like a peer group or a market), shown natively within a wider workflow that enables some click-to-action.
As we extend this list, a fresh set of questions reveal themselves that must be addressed head-on. Questions that one can avoid with a dashboard that sits out by itself. Firstly, we’re asked to check if our analysis and the underlying data are ‘good enough’ to propose an answer to a business decision. This requires us to re-engage in the user journey: to understand their motivations, the context for their decision, and the information to hand.
Next, we consider what is needed for users to truly embrace our analytics: more fuel or less friction? We can overfocus on fuelling adoption, believing that enough fuel can overcome any resistance – like an inefficient swimmer who gets there in the end. Ask the question: is this default mode distracting our attention away from the myriad of factors causing friction in the system? The truth here is often that identifying resistance is just harder, but finding it can be a revelation and removing it can be simpler than you thought. Think again about our hard-working operations colleague. Rather than adding more data/charts and rolling out training programmes, imagine the data visualisation surfaced as a widget inside their existing workflow system, contextualised for the decision they needed to make. The colleague saves time and clicks, and retains the right to exercise autonomy and human judgement.
The business case for embedding analytics here should not be written solely on the basis of our colleague’s productivity. It’s a statement of intent to realign on vision between product and analytics teams, as they weed out organisational inertia and expedite data-driven decisions within the business. And if that’s not enough, the system gets to capture extra context about the business decision made relative to that proposed, the time taken to exercise it, the data available at that moment in time and other features which may prove essential to educating the model and its inventors in future.
The final question is how to set analytics teams on the right path to successfully develop solutions that embed themselves deep in the organisational value chain. We do this by distilling the set of capabilities needed across technology and processes within the firm. By investing ahead and centrally to deliver these, we can create pathways for analytics teams to follow that promise certain characteristics around embedded analytics within their solutions. Some examples follow:
- By establishing a workbench for data analytics discovery, we build a network of like-minded colleagues collaborating on ideas and amplify their voice.
- By making it our business to share certain corporate reference data (like a firm’s product hierarchy), we pull teams closer together and create a hub for those who need to link and aggregate their transactional data.
- By promoting integration of dashboarding environments, we can publish a visualization as a widget in another system.
- By integrating with any enterprise workflow solution, we make it easy for data to pass through analytics solutions, natively and naturally within a business process.
- By integrating with any robotics process automation (RPA) or contact centre solution, we build digital ramps for offline-to-online, and vice versa.
- By establishing common identity and access management technology across the firm, a user’s identity and context can pass between web solutions (built by different teams and hosted in different locations) as they navigate.
- By separating the software development lifecycle from analytics development and content management, changes in one can flow independently and at their own pace.
Our Aon United journey to embedded analytics echoes our commitment to deliver more for our clients. As we reach for higher and higher levels of maturity in our analytics practices, we get to enjoy greater levels of engagement across Aon and reap the rewards of an environment that cultivates new ideas and opportunities. In turn, we continue to invest in yet more shared capabilities within our technology platforms, and further reduce the time-to-value for the next wave of initiatives. And generally, the progress feels real.
Whilst the impact of our analytics often plays out on a broader global stage, there is a strong sense of purpose in working to create technology that works for people and not against them, elevating the expertise that our firm has cultivated in its history.
Karl Heery is VP of technology for data & analytics, RPA & DevOps at Aon