Striking The Balance Between Machines And Humans

by Jack Cooper Andertons Music Co.

Musical instrument retailer, Andertons Music Co., is best known for its stellar online presence with one of the United Kingdom’s top e-commerce sites. With a following of over one million users online Anderton’s Digital Marketing Manager, Jack Cooper, provides insight into how to harness marketing technology whilst remaining influenced by humans.

As a growing e-commerce retailer (UK Top 100), one of our biggest frustrations is the expectation that technology is a one-size-fits-all solution. More often than not, brands are sold the dream, only to later come to the realisation that technology – however comprehensive or automated – requires a certain degree of human interaction. Over time, we’ve learned that the best successes come from striking that fine balance between machines and humans.

For all of the demonstrations, events, and pitches that I have been a part of; these technology integrations scream out for clarity around the actual time and resource investment likely to be needed in order to extract maximum value. And by this, I mean, how much human input is required for these tools to actually provide a clearly improved customer experience. I wanted to share some insight in to how we at Andertons combine emerging tech with our team on the ground to enhance the end solution.

Here are some examples of where we feel technology is great, but ultimately better with a little help from us humans.

Is Automation All It’s Cracked Up To Be?

There are so many things in life that can be automated, marketing included. However, we need to approach with caution and here’s why.

Whilst there are clear instances of automation being a no-brainer, there are plenty of examples when it delivers a sub-par user experience. For example something with a defined period of time, such as ‘It’s been one year since you last made a purchase’, is always true and can help to nurture and retain important customers. Something such as ‘you bought product X, so you should be interested product Y’, becomes subjective and not indicative of typical expected journeys. A couple of unusual transactions are enough to misalign data.

Instead of banking on automated journeys, we create our own, based on our data collection and in-house product knowledge, using technology to implement it. Automation is simply the process of triggering communications. We wanted to deliver on human expectations and we call these experience-led journeys. Drilling down as far as product level, we offer post-purchase content to get the most out of your product, or some further reading before jumping into product recommendations. Something we do to help retain customers and not hard-sell.

Getting Personal Doesn’t Mean Relying On Technology!

Is your personalisation software delivered by a company that doesn’t know your business or have a knowledge of your product? Relying solely on browse and purchase behaviour limits potential.

We’ve set about creating our own personalisation engines, by creating our own rules and weightings driven by our own expert product knowledge, not just looking at shopping signals and black-box algorithms. We input a whole bunch of data, pair it with known shopping behaviours and voila – a basket recommendations widget that is created by product experts, not a computer. We’ve also created what we call ‘pure associations’, where irrespective of the user these products are hyper-relevant to the cart.

We use technology to deliver these, but not to create the headline story. We’ve taken all the things that we believe to be great about recommendations engines, and created one that is bespoke to how our customers like to shop. Not only can we benefit from an uplift in AOV, margin, and items sold, but customers get custom-delivered product recommendations from product experts.

Can Machine Learning Really Super-Charge Your E-Commerce?

Probably, but can it replace it? Certainly not! As the name suggests, machine learning takes time to learn. In all my experiences, machine learning can be fantastic for deep diving into your platform data. A way to make sense of the bottomless numbers derived from channels like SEO or PPC. What it isn’t best-suited to is delivering a bulletproof end-user experience. With a catalogue of over 20,000 products and users that have unique tastes and brand affinity, we wanted to encourage users to give us an insight into the colours, shapes, and sizes of instruments they like as opposed to having to make assumptions based on product categories and less-obvious aesthetic signals. That’s why we set about making Guitar Tinder! An engine to drive our ‘gear you might like’ recommendations.

Creating an interactive digital experience that feeds the learning process (quicker than waiting for users to browse or handing over the data) by helping us to understand what colours, guitar body shapes, finishes, and brands users are interested in.

On-SERP SEO Is A Real Curveball

Search engine result pages (SERPs) have evolved so much with regards to the information they deliver, and one of the least talked-about changes of recent times is on-SERP SEO. Search Engines show so much more information at source than in previous years. Users don’t even need to click a link to your website to get to the content or the answers that they want – essentially turning the very practice of SEO on its head.

When looking at data, consider that CTR is no longer a measure of SERP dominance. Search engines are now a data centre to feed voice search queries and wearable tech. A high number of page one impressions may suffice!

To help bridge the gap, we focus on creating content that aims to boost not just our visibility and traffic from SEO, but also to answer common queries in order to win featured snippets, rich video results, and more. A user intent based acquisition strategy focusing on longer tail queries and questions with high shopping intent. If you haven’t already, it’s time to chase down those intent-based queries. This will no doubt be more prevalent when sites and retailers begin to become eligible to mark-up content for voice search.

Replicating The In-Store Experience Online

There’s so much noise about making the in-store shopping experience more technology driven, more immersive, and using digital channels to drive footfall, but often a lack of conversation around making the in-store experience truly felt online.

As a family business, we value the shopping experience both on and offline, so we created the ability to provide custom quotes for shoppers online. Users can speak to one of our mail order team, bundle selected products together and still use the web interface to organise delivery and checkout.

We also invested in unique photography for our serial-number tracked products. When multiple products are available, shoppers can peruse the actual product that they may go on to buy, and not just a stock photo. This is paramount in our industry where the finish of each guitar is both unique and important. It was essential we used technology to alleviate any online friction.

Key Takeaway

If nothing else, you need to question what you could do to squeeze the most value from OOTB technology – moving away from the set and forget mentality. For all the trailblazing tech solutions, expect to have to customise it to be effective.

At Andertons, we’re fortunate enough to be agile, fluid, buzzword friendly, and most importantly, housed internally with specialists in each discipline. Not only can we break down common barriers to new technology, but we can test and pivot quickly. Whilst technology gets more advanced and humans and algorithms demand a more personalised online experience than ever, is machine learning driving more distance between the two?

Optimise the heavy lifting, but don’t expect technology to be replacing all of our jobs just yet. After all, a successful business will want to retain that customer-first approach to marketing.