burger menu

#YardInsights – Data Science Powered CRO Event

5th November 2019 by Scott McPate | Data

We finally brought our first #YardInsights event to London on the 30 October. Hosting alongside our partners Adobe, at their Customer Experience Centre, we invited four speakers to share their insights on the future of Data Science Powered CRO.

Axel Heyenga from Adobe delved into the future of retail and how the customer experience should be at the heart of what brands are trying to leverage. Stephen Pavlovich from Conversion.com discussed how to use experimentation to help tackle some of the biggest problems your business faces. While Paul Newbury and Emily Davies from Yard explored the use of propensity models and neural network-based scoring that they have been implementing using client data.


Axel Heyenga, Adobe
Future of retail

Our first talk of the day came from Axel, highlighting that experience is one of the most important aspects that brands and organisations should be focusing on. While digital has changed the retail landscape, there should now be greater importance on driving emotional connections to the brand through unique events, pop-ups or store experiences. Then from this, you collect data through enhanced personable connections.

He went on to explain that we need to connect data, content and the store experience into one journey. Being connected with your customer is not a new idea! Don’t think about technology as the first port of call, you need to be adaptive around where the customers are and predict where they are going. Importantly experience and loyalty define how retailers perform.

ABCD of Experience

  • Give customers an adaptive experience.
  • Be wherever they love to be.
  • Help them filter choices.
  • Differentiate through immersive experiences.

Axel explained that retailers have to respond in a few different ways. The need to redesign the retail experience by creating more captivating online and physical stores. The customer journey needs to be optimised across screens and locations. Personalisation has to be scaled by leveraging AI and ML to anticipate desires. And finally, establish an endless shelf with the digital portfolio to be adaptive.

And remember:

“Last best experience becomes your minimum future expectation.”


Stephen Pavlovich, Conversion.com
Using experimentation to drive product

With lots of experience of working with clients that are looking to innovate with product, Stephen highlighted some of Conversion.com tips for experimentation.

  1. Use experimentation to solve your biggest problem. If you want to avoid frustrating all customers, experiment on a small basis of customers to help see how this will impact results.
  2. Be bold. Most decisions are reversible and you may fail or you may succeed, but experiment depending on the scale of your organisation. If we aren’t prepared to fail, we can’t innovate.
  3. Test early and test often. With dendritic branching, you can constantly test, validate and iterate to explore new ideas.
  4. Start small and scale. Prepare to truly innovate at this stage between iterations and try to spark the imagination of people on ‘what could be’ and not ‘what is expected’.
  5. Measure what matters. Adapt your evaluation criteria as you iterate on experiments. Spend time and money testing on a small group of customers and assess their reaction rather than roll out changes without testing.

Ultimately this approach will provide you with clearer evidence on what works and takes you to the point of becoming a product visionary. The experimentation method has a low-cost investment attached to it and allows you to test more ideas, more frequently. It also means you can test high-risk ideas with a safety net and provide autonomy to teams.


Paul Newbury and Emily Davies, Yard
Propensity Modelling: Predicting the Customer’s Next Move

The focus for Paul and Emily’s talks centred on the multiple interactions customers have with organisations through several different channels. They pointed out that visitor records need to be joined into a single visitor in order to get a view on the real value of each touchpoint. Customers want more than just a product or service they want an enriching and smooth experience across all touchpoints.

To help identify opportunities at Yard we use propensity modelling. This is a way for us to identify who among your audience is most likely to make a purchase, accept an offer, or sign up for a service. This approach can be used to predict negative events in order to optimise activities and identify when the customer is likely to churn and take preventative action.

Paul and Emily explained that the Yard model uses around 120 input features along with built-in hyperparameter tuning. This ensures we can build the best fit for every product across every client. It’s also better to use modelling in real-time allowing instant decisions to be made, rather than using it at the final stage of a business process.

Machine Learning is not magic – it provides additional insight and data points that are hopefully helpful to allow us to offer a better customer experience.

With all of these input features, Yard have to utilise this data into workable experiments and optimisation. In most cases, we can tie in specific content targeting, matched with consistent messaging and this can help increase ROI through customised messaging and repeat engagement.

 


If you would like to join us for a future event or would like to receive a copy of the presentation decks then please get in touch with us hello@yarddigital.com

Share this post

Related Content

The Back Yard – November 2018

Yard – Diversity of Thought in The Workplace

For more insights into agency life, subscribe to our newsletter for the latest Yard updates.

The Edinburgh Festival Fringe 2017

  You’re surrounded by crowds on Princes Street, there are street food stands popping up all over the place and in the space of five minutes you’ve walked past...