When is it the right time to hire a data scientist... and why does your ecommerce business need one? Before we answer those questions, let's take a look at what the data scientist does and the skills that he or she needs to have.
The simple yet effective equation for a Data Scientist is essentially a balance of understanding statistical maths, computer science technologies, database design knowledge, business acumen related to a particular domain and interpersonal communication skills to transform and modify businesses for the better.
If you think that definition sounds a bit complex and overwhelming, you’d be correct. If you think there’s a lack of individuals in the market who would be able to tick all those boxes, you’re right. If you think they’d be able to add immeasurable and infinite insights to your business…well, you get the idea. Without data scientists and until you apply data as evidence, decisions made in the business are not smart. It’s not enough to go with your gut feeling anymore. Reviewing the statistics before risking a new product launch or a marketing campaign just makes sense.
What can a data scientist do for your brand?
Data Scientists help to turn brands into meaningful brands, and strengthen relationships between themselves and consumers, constantly adapting to the needs of the consumer so their brand is not only trusted but also valued to the majority of the population.(1) In my experience, most of the data analysts I speak with from the ecommerce industry are missing a Data Scientist from their wider analytics team and it makes me so disheartened but also not surprised to discover the gap is still so significant in these departments.
Retailers who currently either employ an in-house or agency-supported data science capability in relation to their ecommerce offering, as well as many other areas of the company, include John Lewis, Farfetch, Moneysupermarket.com Amazon, M&S, Aldi, Lidl, Paypal, Sony, Sainsbury’s, Samsung and Boots. Even if these examples (even with their widely spread sub-industries) don’t necessarily filter in to your idea of a direct competitor, they’re still incredibly robust case studies to aspire to imitate.
What I mean to say is, whatever your area of focus for 2017, your business will require the services of an amplified data analyst – one who not only is aware of and makes meaningful use of the historical and previous data, but has the technical skill of coding and programming to build algorithms which record live data and can predict trends before they happen, thus retaining your loyal customer base, but also engaging with the right audiences to target new customers.
The main ways they can improve the strategy are applicable to all ecommerce business, but below are the top five areas of where their efforts are best utilised in a bit more detail. These include identifying smaller shifts happening in the industry, shortening the marketing lifecycle to improve personalised marketing strategies, predicting the supply chain model, broadening the customer base and defining the product mix to personalise presentation of the products for each site visitor.
5 ways a data scientist can help
1. Identify smaller shifts happening in the industry
…to stay ahead of the curve when retaining their core customers. Data Scientists do this through developing powerful algorithms linked to innovations enforced by competitive retailers in real time, continuously monitoring the landscape with cross-referenced external sources (or web-scraping) to find trends.
This is a lot more effective than manual examination and as a consequence can also influence an imitation amongst customers in regards to shopping habits, building engagement levels but also brand loyalty. The more you know your customers and mould your offering to their preferences before they even know they’re preferences, the more likely they are to return to your website regularly.
A case study of this would be the development of TwitterReverb, a backend application which allows users to reveal and visualise patterns in Twitter activities surrounding particular brand hashtags, topics and keywords, monitoring the active responses and the content they want to retweet, thus furthering the communication (2)
2. Shorten the marketing lifecycle and improve personalised marketing strategies
...to attract more customers and reassure current customers to make more purchases. Ad retargeting, ad word buying and channel mix optimisation are just a few ways Data Scientists do this, again, by designing algorithms to launch tests and experiments relating to various strategies.
They’re able to understand what factors led to the purchase by analysing the events occurring in the lives of the customer beyond regular shopping habits by using graph analysis on social networks, so you not only have the blind faith but also concrete evidence to spread the budget wisely.
A real-world example of this would be the Adidas/Reebok affiliate programme, bringing all the independently operating markets into one collective streamlined operation and expanding this into new markets. The data insight collected in this controlled strategy involved matching relevant demographics to specific promotions and led to a significant three-digit percentage revenue growth and an average of 33.5% decrease in cost of sale for both companies. (3)
3. Predict the supply chain model
…for more effective and engaging delivery options which suit your audience. Organising the correct amount of products in the right place at the right time is made more seamless and effortlessly aligned with the assistance of Data Scientists, who are able to develop advanced predictive models which diminish risk on the equilibrium of supply and demand.
This is ultimately instrumental in satisfying the customer journey and enhancing profitability relating to product sales, informing strategy around trading for future campaigns.
Amazon is working with data scientists to predict the supply chain model with a pending patent called anticipatory selling, which allows them to gather information on past orders from particular customers and build a prediction model which uses this data and graph analysis to identify what their regular buys are.
Constantly taking it to the next level, when this is launched it will mean that the product is dispatched in advance to satisfy the expectation of delivery times once the purchase is made, using the time whilst the product in in transit in the local geographical area to promote the product through online marketing efforts.
4. Broaden your customer network in gentler ways
…to reach social media influences in your current customers in order to obtain new ones. Data Scientists have the ability to monitor the social channels where your customers have the opportunity to actively promote the brand to their network through reviews.
They observe who is responding to their activity through comments or likes, whether that be friends, family or colleagues, and then storing these information trails as data. Using this insight can help a retailer to understand what customers want and improve accordingly, or target those connections in more passive ways.
For example, building a network graph for a social media platform regarding the Twitter handle for @Walmart in America, showed a strong tweet connection between the company and one of their most shared products, Huggies nappies. Studying interactions in the communities in which your customers interact can allow you to take advantage and action of this. (4)
5. Define the product strategy for ultimate product mix presentation
…to forecast the opportunities in production. Data Scientists build algorithms which can help e-tailers identify and adjust the product mix by discovering where the gaps are in the current season, what products should be made and how many should be ordered, what time (up until the hour on the day) they should be advertising the products to sell and when to cease the supply of a certain product entirely.
By implementing more cutting-edge technology to design predictive analytics platforms, you can ensure your data is lightyears ahead of your competitors, also offering a bespoke service to your customers through understanding what filters through which they shop on your website to cater to their preferences, integrating cross-sell and up-sell options more fluidly.
George @ Asda, the supermarket’s clothing range, recently were able to adapt this to multiple devices where customers browse and shop their site by redesigning their mobile format to incorporate their implementation of product mix modelling for a more targeted and thus effective shopping experience.(5)
How to Hire a Data Scientist
Data Scientists can add value to ecommerce business through more advanced analytics and this can only become more sophisticated through the channels through which your customers shop and the revolutionary technology available to them.
With more demand than supply of high-quality, bright and digitally aware Data Scientists out in the market at the moment, it would be advised to make 2017 the year of the hire and engage right now with a limited, niche community who can really make an effective contribution to not only the general profitability and customer satisfaction, but more importantly, the overall strategy of the business and multiple teams.
You may already have Data Scientists who are moonlighting as Data Analysts, or haven’t defined their potential as yet. Have you individuals who are able to contribute this level of recommendation which are labelled as something else? There may be a new title which may not reveal a discouraging “Sorry, no results containing your search term were found” outcome of which I’m unaware!
Appendix/Sources: http://www.havasmedia.co.uk/meaningful-brands/ http://www.periscopic.com/our-work/visualizing-how-conversations-unfold-on-twitter https://www.iprospect.com/en/gb/our-work/adidas/ http://www.datasciencecentral.com/profiles/blogs/visualize-your-social-media-analytics https://www.iprospect.com/en/gb/our-work/asda-george/