RETAIL NEXT BEST OFFER
What if you knew what each of your customers wanted, when they wanted it and via which channel? The ability to then provide tailored, timely offers would increase exponentially; the impact on your loyalty program would be substantial. By combining all of your customer interactions – physical, digital, emotional, transactional - into a single view record, our next best offer option can determine the behavioral patterns of individual consumers, opening the door for a truly customized experience. You can use future insight to inform next most likely action at the individual customer level.
Our next best offer scoring engine boasts 90+% accuracy on what your customer is most likely to want next. You can then focus on developing the right offer and providing it to them at the right time via their preferred channel. With this micro-targeted view of customers, organizations can systematically predict, match and provide effective offers that will: enhance brand loyalty; see a higher return on marketing investment; real sales uplift; reduced churn; and improved shareholder value.
CYBER CRIME DETECTION
Cyber crime is now well-funded, organized and is most often committed through unknown attack vectors. The USA, Australia and UK are the top three ‘attacked’ countries in the world and incidences are increasing annually. Attackers are increasingly sophisticated and for many organizations, just keeping up is difficult.
Working with the specialist EY Cyber team, EYC3 develops purpose-built cyber protection recommendations using our advanced analytics platform, EY HEX. Our system can ingest more than 10Tb+ of network data each day and we apply behavioral analytics to network assets (such as IP addresses) searching for abnormal behavioral markers. We identify what an attack (kill) chain looks like and then monitor and detect such intrusions at scale. We ensure our algorithms are continuously learning and adapting. There is also the ability to incorporate automated intervention and end-user control.
We are able to implement a pilot in weeks, not months and we have proven that we can identify previously unknown attacks, in some cases as many as 30 per month. We develop the cyber attack use cases in consultation with our clients covering areas such as: process anomalies; network recon and lateral movement; unauthorized access; and data staging and exfiltration.
PREDICTING MASS POPULATION BEHAVIOR
We use the ‘how’ and ‘why’ of decision making to identify change or trends before they occur with minimal data. Our mass population modelling is based on the premise that people’s behavior changes but their cognitive decision processes do not. Behavior change is driven by changing context and influences and by focusing on how and why people make decisions (the cognitive decision process), we’re able to accurately predict future behavior across large populations using minimal, readily available data. Our team uses a multi-disciplinary approach with experience across artificial intelligence, simulation, behavioral science, game theory, agent-based modelling, engineering and micro-economics.
- Water case study: Our on-demand water model has been used by a number of clients globally to forecast residential and non-residential water usage into the future. This has included testing the impact of water conservation measures; analyzing school usage behavior to identify and quantify the impact of leaks and overuse of utilities; the bounce back effect on water demand once droughts end; future pricing & incentive models; and water demand and ‘what-if’ scenarios across customer bases.
- Energy case study: EYC3’s world-leading predictive residential electricity demand micro-simulation solution was used to model the impact of disruptive factors (e.g. electric vehicles, solar, battery storage) on future residential energy demand. It was able to forecast future residential electricity demand across Australia, with regional breakdown, given future disruptive factors, including appliance/technologies uptake, and changing climate and demographics.
GROW AND MONETIZE THE FAN BASE IN SPORT
EYC3 has experience in advanced sports analytics, specifically in road cycling and football. Using machine learning techniques, we have undertaken a number of predictive and prescriptive analytics projects to: predict race/match outcomes; run test & learn scenarios; and create virtual race/match simulations to understand the impact of random factors on outcomes.
- 2015 AFL Grand Final winner and winning margin predictions: we predict the winner and winning margin for the 2015 AFL Grand Final match between Hawthorn and the West Coast Eagles. See how we did it and our analysis of the final results.
- 2015 Tour de France virtual road race: using stage 20 of last year's Tour de France as a backdrop, the EYC3 data scientists used predictive machine learning analytics across the International Cycling Executives (ICE) member’s Strava data - more than 1 billion data points - and constructed individual ICE rider “power curves” to predict the yellow, green and polka dot jersey winners. Watch behind-the-scenes to learn more about how EYC3 created this virtual race.