Shopping Analytics: In order to know your customer, observing is not enough

“Know your customer” – probably one of the oldest wisdoms of the business playbook, but often easier said than done. Image you are the owner of a local grocery store, having a product range from fresh fruits and cheese, to toilet paper and toothpaste. At the end of each day, you receive a set of receipts, detailing each purchase with information such as products bought, basket size, time of purchase and method of payment. Making substantial sense of this data with actual decision impact is already a tough undertaking: Customers purchase more alcoholic beverages after 6pm, so should they be placed more prominently in the evening? How can customer behaviour be reliably analyzed and customer motivation be leveraged for improved sales?

Validating certain hypothesis about customer behaviour is significantly easier in e-commerce. Changing the arrangement and placement of products in an online shop and immediately measuring the impact on user activity is easy and common practice. A/B testing, which allows for rapid iteration and optimization, is rigorously applied. To be fair, supermarkets have found feasible ways to also improve their data availability. Customer cards – the most notable example stems from UK-retail giant Tesco – allow supermarkets to track the purchasing history of individual clients, segment and understand their target group better and develop customized product offering. Additionally, the rise of the smartphone has made sophisticated analysis of customer behaviour in supermarkets and retail in general easier. Information about purchases is now complemented with data about how customers navigate through the store, where and how long they stop at each aisle.

On macro level, data-driven decisions about store design, optimization and arrangement can credibly be taken. Nevertheless, while information about customer behaviour seems readily available, the question in offline-retail often remains: Who is actually shopping? What is their personality? Beyond some generic information about age and address, little is available.

This question is far easier to answer for those retailers, which allow their customers to conduct their purchase with a smartphone application. As in-app shopping is rapidly growing and most major retailers have opened up this additional channel. It allows for the highest level of personalization given the amount of data, which can be aggregated from other sources – especially social media and GPS data, which are all associated with the owner of the respective smartphone. Nevertheless, as information has been exponentially growing across channels, one question remains: What motivates customers to purchase a certain set of products?

Even as data availability, volume and methods of integration and analytics improve, the gap between understanding behaviour and understanding motivation has not been closed. In the offline world, traditional methods of market research through interviews, focus groups and surveys still complement methods of big data analytics and – even though their importance slowly diminished over the last few years in the opinion of a few marketing experts – will probably continue to do so in the upcoming years. Nevertheless, smartphones, as the daily companion in the digital age, are a strong asset for market researchers, as they allow for rapid interaction with and feedback from the target group. However, in order help make valuable decisions about consumer motivation and preferences – going beyond mere behavioral observation – three things need to be taken into consideration when thinking about the integration of targeted survey options into shopping apps.

Limited methodology options

While surveys can easily be implemented into native apps – for example with tools like Apptentive – companies should clearly establish the goals of their survey. As opposed to real-life surveys, where respondents focus deliberately on the given question, the smartphone as a medium is crowded with other applications, exposed to a constant stream of push-notifications. User attention span is limited, as are methodology option. Given the size of a regular smartphone, multiple choice questions can display a maximum of 4-6 options. Open questions, which require unique answers are certainly possible, but also not too suitable for the medium. With this limited set of methodology options at hand, a clear-set objective is imperative. If surveys should go beyond a mere feedback channel with direct customer interaction, data quality needs to be a primer.

UX is crucially important

In order to supplement the broad array of data analytics in a meaningful way, companies should maintain attention on response rates, because certain thresholds need to be surpassed in order to be able to draw meaningful conclusions from a data perspective. Any e-commerce app naturally focusses on a unique and seamless shopping experience. Also when deciding to gain more insights into customer motivation through integrated surveys, companies need to make user experience a top priority. Should the survey look like a survey? Or could it be integrated in such a way as to be barely noticeable? Some apps – for example the fashion e-commerce app Swipey – almost work as a constant survey. Swiping through different fashion items, users are “forced” to make a deliberate and active decision regarding certain products. In such a framework, new products can easily get valid feedback before launch day. Similarly, market research company Nielsen had already 2013 launched an app – officially branded as a “Top 10” app to help users navigate trends in books, songs and TV shows – whose sole purpose it was, to constantly survey its users for opinions.

Impact on conversion rates

Executives pondering about whether to utilize the app’s user base for the generation of new attitudinal data, should always bear in mind: Anything away from the core business purpose of the company (in this case: selling products), could potentially drive down conversion rates and thus have a direct negative effect on sales. On the contrary, implemented surveys could certainly be the key knowledge source for further app optimization, product customization and thus revenue growth in the medium-term. Given this delicate balance, executives are well off to first initiate an intensive discussion with the company’s data team. Generating any kind of additional data – on top of the already existing bulk – is only a valid choice with a good and promising use case.

Helping to bridge the gap between customer behaviour and motivation and thus enhancing the shopping experience and revenue prospects is a key endeavour in on- and offline retail. With the vast amount of data available in smartphone-based e-commerce, one seems to be tempted to include survey options to generate additional dynamic, high quality and highly personalized data. But oftentimes, knowledge potential on existing data analytics in not even fully tapped. Before thinking about implementing survey options – with the described challenges -, executives should always consult with the data team in order to develop clear-cut hypotheses and use cases, which add serious value beyond conventional big data analytics.

Big Data

Leave a Comment

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.