Please enable javascript in your browser to view this site!

Applying Data to Improve CX

Personalisation emerged as the key theme of October’s panel discussion, Applying Data to Improve the Customer Experience. It has toppled segmentation as the go-to marketing model for defining audiences, and is derived from the increasing amount of consumer data available to businesses. Personalisation is also challenging business - and agency - self-perceptions of customer centricity, exposing pitfalls in established processes and supplier relationships that constrain a truly customer-first approach.

We explored

  • The differences between segmentation and personalisation, and the influence of data in this transition

  • What it truly means to be customer-centric and how to obtain meaningful insights

  • Established structures and beliefs that may be holding businesses and agencies back

  • The role of technology

  • The ethics of personalisation - the use and application of individual data

  • The future of the workforce - what skills will be relevant? What should we be upskilling in now?

With panelists

Edward Crouch, General Manager - Lavender
Nigel Dalton, Chief Inventor - REA Group
Mark Chatterton, Cofounder - inGenious AI

Targeting your market, a brief history: Segmentation vs Personalisation

As manufacturing and market size increased in the 1920s, segmentation enabled businesses to highlight relevant value points. Segments, or models, were based on demographic, socioeconomic, lifestyle and psychographic factors. Hyper-segmentation has allowed us to further narrow these segments, and with Human Centred Design (HCD) came the idea that communities (or segments) could be represented by a single coherent identity - a persona.

Personas have dominated the way campaigns have been developed, tested and executed since OglivyOne adopted it internationally for their clients in the 90’s. While the earlier applications of persona’s required consumer research and data, out of the box identities have become the time and cost efficient option. The Millennial, The Baby-Boomer, The Yuppie have all become stereotypical examples of off-the-shelf consumer research where publicly available statistics, or assumptions, have been used to develop brand-specific communications.

The advent of digital has meant businesses have more access to information about their customers than ever before. They can track it, store it, sort it and use it to personalise their communications. It is the next evolution of the original intent of segmentation - using specific demographic, psychographic, socioeconomic and lifestyle factors, alongside of individual interactions and behaviours with a product or service.

In a perfect world, technology - while still evolving - allows brands to create individual experiences and interactions with their customers online and offline. Target can identify a woman is pregnant based off her purchases and market relevant products, Spotify’s ‘Discover Weekly’ brings its users curated playlists, and Delta Airlines launched a handheld device giving flight attendants access to passenger preferences and relevant information. This move toward individualisation better positions brands to build affinity, and raise the bar of customer service and experience.

Say it’s not a perfect world - how can personalisation be achieved in reality?

Many businesses simply don’t have the resources of Target, Spotify or Delta, and while they may have some data, understanding the most efficient way to access, interrogate and utilise it poses challenges.

Personalisation is really symptomatic of a customer-centric business. While many claim to be this, they often lack the data, structure or both to truly realise it. REA Group have approached the challenge in two ways - through process and research. Their structure is now based in lean and agile methodology, allowing analysts to go where the work is being done (i.e. customer service) and base their solutions in tangible experience. This approach has removed the traditional barriers that separate silos within a business, preventing the sharing of knowledge (and data!). REA Group’s research methodology is based on one simple directive - just talk to people. A room in their head offices was repurposed to be a Consumer Research Lab, where they have customers in each day to talk.

Face-to-face interactions with the consumer’s you are trying to reach can provide invaluable insights. For REA Group, it is one powerful statement from a consumer that informs much of their approach - “Thank God we found a house, now I have my life back”.

Talking to your customers can also help you better understand how your current interactions are measuring up. You can’t achieve positive and meaningful interactions with customers if you’re only looking at your output - seek feedback, seek it regularly, and seek it face-to-face. Investing in this kind of research and these conversations does pay off. Clare-Marie Karat from the IBM Watson Research Centre referenced a study in Cost-Justifying Usability reporting that $20,700 spent on usability resulted in a $47,000 return on day of implementation, and in Software Engineering: A Practitioner’s Approach Robert Pressman demonstrated that for every dollar spent resolving a problem in product design, $10 would have be spent on the same problem in development or $100 after release.

What does personalisation mean for agencies?

Agencies lifeblood is the retainers and projects of the clients they work with, ideally aiming to be the go-to for strategy, production and execution. Personalisation has meant that the volume of work expected by their clients has increased exponentially. Agency pricing structures and processes are not set up to meet this demand, and as a result many businesses are bringing skills in-house by hiring their own designers, or buying out their suppliers; for example this year Swisse Wellness acquired the controlling shares of their advertising agency, Noisy Beast.

Personalisation is forcing the full-service model out, to make way for consultancy. Indeed, established consultancies have been acquiring independent agencies; in 2015 Accenture bought award-winning digital agency Reactive, and followed it up this year by acquiring creative and design agencies The Monkey’s and Maud to expand its customer experience capabilities. The transition from full-service to consultancy is shifting the value-proposition of suppliers from production to analysis and solution design. To move with the market, agencies can no longer be everything to everyone - they have to go niche.

How is technology like AI using data to improve the Customer Experience?

As we learned from September’s panel discussion, the use of data in AI and Machine Learning (ML) is driving personalisation to new levels, manifesting most commonly in recommendations like the Spotify ‘Discover Weekly’, or Chatbots.

The National Bowel Screening Program offers each Australian between the ages of 50 and 74 free bowel screening every two years. Individuals are sent a testing kit that can be done at home before sending their samples to a pathology lab, however in many instances tests are not being completed correctly. Incorrect testing can lead to higher costs and unnecessary additional work for health professionals and individuals, extending an already uncomfortable process.

Work is currently in development on a chatbot where individuals receive timely reminders at each stage of the pre-test process and can ask basic questions, aiming to reduce the number of incorrect submissions. The technology is providing a safe environment where recipients can ask questions they may be too embarrassed to ask a nurse or doctor, and similarly the pressure is taken off health professionals by having simpler enquiries managed for them. The same approach is being developed for health insurers where ML is assisting individuals build their own cover from over 9,000 possible cover combinations. When a policy is determined, customers have the option of seeking further clarification from a customer service agent.

In both of these examples technology is used to do the grunt work, refining the conversation for humans further down the line. Their value proposition to the organisation and individual is clear, and their primary objective is service. Where AI and ML go wrong is when the value proposition is not adequately defined and the objective to sell overrides service and ethics.

How do humans fit into the data/technology equation, and what are the skills we should be focusing on in this environment?

Humans are the ultimate source of insights, data and solutions. The interface in development for the National Bowel Screening Program was made possible by the insights and information provided by health professionals, and while AI is doing amazing things with the input of data, the output still leaves something to be desired - more complex questions and problems require empathy, and we can’t build an algorithm for this (yet!).

While AI and ML are emerging, and the technology and titles that come with it seem intimidating, the skills to be successful in this field don’t necessarily require you to invest in upskilling. For example, a Conversation Interface Designer might have begun their career in Human Resources. Consider these three key roles and their associated skills in developing data-based solutions and technology for CX;

  • Analysts: No matter how advanced the technology, or granular the data, you still need to correctly identify and define the problem or need.

  • Data scientists: While ML takes care of sifting and sorting through data, it does so according to a framework that has been built to match the problem or need. This is done by a data scientist where the study of actuarial science (mathematics) is crucial.

  • Communicators: You may have the designed the best solution in the world, but if you can’t explain it to stakeholders then you haven’t succeeded.

Beware: cautionary tales of personalisation in customer experience

Data, our access to it, use and application of it has created ethical dilemmas for businesses. While Target were able to identify and disseminate relevant product information to pregnant customers using data, the execution garnered negative publicity for the brand when it unwittingly outed a young pregnant woman to her father. Similarly, (though less dramatically) Amazon were condemned for automatically adding recommended products to their customers baskets, and many of us are now well aware of the costly pitfalls of searching and buying flights online. The key refrain for the night was, “don’t be creepy”. Don’t take the decision, or free will, out of the hands of the customer, and if you cross ethical lines in the use of personal data then expect to exposed.

It is important to note that in a world where betting companies are using data to provide personalised recommendations for future bets, and order upsizes are increasing at McDonalds with the introduction of touch-screen ordering systems, our boundaries are being tested. Education and the increased commonality of AI and ML is in turn increasing individuals comfort with this technology. This allows them to take advantage of positive applications like bowel-screening chatbots, but it may also be increasing our tolerance for ‘creepiness’.

Nigel’s Top Reads

Nigel Dalton is clearly a well-read gent, and we benefited from many recommendations made throughout the night. These included;

  • Yuval Noah Harari, Homo Deus, A brief history of tomorrow

  • Kathy Sierra, Badass: Making Users Awesome

  • Steve Krug, Rocket Surgery Made Easy

  • Steve Krug, Don’t Make Me Think