Big Data Can Be Too Much and Not Enough

big dataLike many other businesses, today’s multichannel retailers are keen to capitalize on the promise of big data. How can you tap into the increasingly large data sets being generated — both online and in the store — to make better and faster decisions, increase your understanding of shopper behavior and ultimately improve the bottom line?

While it continues to get easier to collect more and more data from e-commerce clickstreams, POS systems, and social networks, many retailers are struggling with how to process and transform all this raw information into useful, actionable insight.

And more does not necessarily mean good if data required for critical answers is still missing. For example, retailers won’t be able to understand multichannel shopping behavior if they only have data about online activity, while remaining blind about their largest sales channel: the brick and mortar store.

Here are several considerations key to addressing the dual challenges of too much and not enough data, and for capturing the insight needed to thrive in multichannel retail:

Avoid data exhaust

Raw data is of limited use. It’s what you do with it that counts. To avoid data exhaust, retailers need to think strategically and be selective about what to capture, so that they can separate the important intelligence from information that’s just noise.

You don’t want to waste time analyzing data that won’t yield anything actionable. Take the time to clearly define what you need to measure (for example, sales conversions or actions prompted by a specific offer or promotion) and then consider which technologies and sources of information will help you understand the metrics that will truly impact the direction of your business.

Don’t miss out on in-store data

In-store data is critical to understanding the multichannel shopper, considering that physical stores are where the lion’s share of sales still happen.

Gartner reports that, even allowing for the increase in revenue from the online channel, retailers should expect 85 percent of their revenues to come from physical stores through 2016.

In-store analytics provide insights into where customers spend most of their time in the store, how they respond to product displays and how long they’ll wait in line before abandoning their cart. And in brick-and-mortar environments, this requires the deployment of technology to collect the data.

Unlike the digital trails they leave when visiting ecommerce sites, real world shoppers don’t click on the door when they enter a store or pick up an item from a display. And while point-of-sale (POS) transaction data from registers has been traditionally used by retailers to analyze in-store performance, it’s use is limited because it can only tell you about the people that bought something, not those that visited but left empty-handed.

In-store technologies are designed to collect customer behavior data at different points throughout the store — including entrances and exits, at queues, in aisles and at product displays — providing a great way to close the brick and mortar insight gap. With more sophisticated options to choose from, multichannel retailers are now making in-store analytics a high priority, and are using the intelligence gained to direct their marketing, operations, staffing and merchandising strategies.

Good decisions require good data

Big or small, data means nothing if it’s not accurate. Many of the methods most commonly used by retailers to collect data about what’s going on in stores involve people with clipboards and visits by mystery shoppers.

These methods are inconsistent because they collect data only occasionally, and also suffer from observer bias and fatigue. And POS systems alone, as noted above, only keep track of shoppers that purchased something and provide no insights about shoppers that don’t buy.

First, accurate traffic data (i.e., people counting) is a must, because understanding traffic patterns in real-time and over time is essential to making decisions about staffing. Used in combination with POS data, traffic data is also required to calculate “sales conversion” (the percentage of store visitors who bought something), which is a key performance measure.

Also important is the ability to analyze customer behavior throughout the store (at which displays do they stop? How long do they spend looking at products? etc.), as this kind of information provides insight into the effectiveness of displays and store layout. And finally, measuring service time, customer/staff ratios and queue wait times helps retailers understand if customers are receiving the service level required.

If any of the types of data mentioned above are inaccurate, the analysis and decisions made based on the information will also be wrong. For example, if traffic counts are off because your data capture technology cannot correctly identify single shoppers as well as shopping units (people shopping together), then you may over or under staff the store. Accuracy is also impacted by the fact that collecting data in real-world environments is more variable and susceptible to error than tracking data online.

When evaluating in-store analytics solutions, there are a number of factors to consider, including the following: Can the technology distinguish between adults and children? Between humans and inanimate objects like carts? Is it able to tell whether a shopper is exiting or entering a location? Accuracy is step one in transforming data into truly valuable knowledge and insight.

Before multichannel retailers can begin to explore the promise of big data, they need to be prepared to not only manage the volume, but also solve the trickier problem of capturing “missing” in-store data that, if ignored, will ultimately lead to missing service, customers and sales.

An ability to zero in on the most relevant and accurate data is key, both online and in brick-and-mortar environments. As innovations in analytics continue to evolve, multichannel retailers will realize new opportunities to improve their business with big data.

Ralph Crabtree is CTO and co-founder of in-store analytics firm Brickstream.