In 2009, Starbucks shuttered 300 stores owing to a 70 per cent slump in quarterly profits amounting to $64.3M. The drop in sales was blamed on the economic slowdown and lower spending power of customers. It only made sense that as personal budgets became more frugal, indulgences were the first to get cut.
Starbucks’ leadership had no choice but to reanalyse the way they did business. Their plans included staying top of the mind with users during the lean period and preparing for a future when customers would return.
Soon, the company realised it was sitting on a treasure trove of data collected from users at their physical stores. Not only did customers spend a considerable time at cafes, they would also sign in to the WiFi. Starbucks built a strategy that aimed at blending offline and online experiences. Online, users could explore content about their favourite blends, track promotions at nearby stores and share coffee lovers through its social media banter. In-store, users were given access to exclusive content that included free news, films, videos, e-books, and music. The coffee brand further merged the two worlds by offering various mobile payment avenues and online loyalty cards, e-gifts and barcodes that could be scanned at point of sale to buy their products. The Mobile Order and Pay feature allowed customers to order on mobile and then pick up their products from a physical store nearby.
Also Read: Get Your Data Machine Learning Ready
Starbucks may have been ahead in the game at the time. Since the pandemic, the line between offline and online has become increasingly blurry. The new expectation is that brands must merge in-store and online experiences in a seamless journey that recognises customers, continues conversations where they left off and engage them with personalised offerings. The data onboarding exercise crumbles data silos and brings all the data about customers into a unified CRM. Marketers will also get more clarity on whether their digital ad spend is reaching the right audience and leading to offline conversions.
But first, brands need to capitalise on the data they collect from their offline sources, ensure it is trustworthy and make it malleable to run machine learning algorithms that can draw valuable insights. Data onboarding refers to the ability to bring offline data online. It promises to dramatically improve cross-channel marketing insights and attribution, expand target audience size, and enhance campaign performance. Think of it as adding a new dimension to the existing data that personifies customer profiles. After all, customers don’t solely exist in online or offline spaces alone.
Typically, brands collect tons of offline data from events, transactions, demos, and data from support desks and call centres. Data from these offline customer and prospect interactions add colour and dimension to make customer profiles more holistic.
The challenge lies in the fact that raw customer data comes in all types of formats and from various external sources. It needs to be standardised and cleansed to fit the requirements of its end destination. This is called data preparation. The essential elements of an effective onboarding strategy include gathering, organising or preparation, uploading and validating data. Automation is the easiest and most efficient way to get this done. Marketers are now choosing data preparation platforms for their data onboarding efforts. Data preparation platforms transform the entire data onboarding process to exemplify the iterative nature of data onboarding, instead of the largely trial-and-error based process under legacy tools. Oracle’s OnRamp, Trifacta, GoldenSource, LiveRamp, FullContact, Adeptia’s Connect are all onboarding solution providers.
Basics of onboarding include structuring the data, which may be stored in different formats like APIs, JSON files or weblogs so that models can make sense of it without causing glitches. Establishing the accuracy of data is probably the most important piece of this puzzle. This includes identifying gaps, mismatched data points and value distribution anomalies which can mislead the model to result in false insights. Cross-referencing data to match online profiles to offline inputs helps build a consolidated view for marketers and other stakeholders. There are solutions that perform functions like joining datasets, pivoting and unpivoting data, aggregating values and deriving vital indicators.
It may seem like a lot of work, but it’s definitely worth the effort. According to an Accenture report in 2017, 88 per cent of potential buyers browse for a product online before purchasing it from a physical store. Online shoppers don’t want purely online experiences because they miss physically handling products. But they also don’t want purely offline experiences because that limits how they can shop. Their top priority is convenience. And it’s up to marketers to make that happen.