h as Kmart have faced consumer criticism for limiting customers on their sites and struggling to keep up with deliveries. Even as restrictions loosen and stores begin to reopen, some retailers, such as Accent Group, have flagged a permanent shift away from bricks-and-mortar towards a greater focus on omnichannel offers. It’s a good time for retailers to consider the value of omnichannel strategies, particularly given the large and continued investments required to do it well. It is undoubtedly crucial in the current market that retailers have a presence across multiple channels. Consumers are increasingly engaging across them and often simultaneously. A consumer might be inspired by a social media post on their phone while watching TV, then search a retailer’s website to see what’s on offer – all at the same time. It is important that brands can engage where the consumers are, and not expect consumers to come to them. Research also suggests that engaging with consumers across multiple channels can increase customer satisfaction – and even loyalty. This all means that retailers now are at a competitive disadvantage without a strong omnichannel strategy. One of the more intriguing questions is how omnichannel strategies impact consumer purchasing behaviour. Glen Senk, the former CEO of Urban Outfitters, was famously quoted as saying, “Multichannel customers spend two to three times more than single-channel shoppers… Consumers who engage with the company across three or more channels spend six times more than the average customer.” This is a consistent finding across multiple industry studies as well as academic research; on average, consumers who interact with a brand across multiple channels spend substantially more than those who only interact with a single channel. For example, a consumer who visits a retailer’s website as well as stores typically spends more than someone who sticks to either channel alone. This is an intriguing phenomenon as it could be interpreted to mean that the act of engaging with multiple channels leads consumers to spend more. If this is the case, then retailers should encourage consumers to engage with more channels, particularly those that have the largest positive impact on spending, a process known as “right-channelling”. However, there is an important alternative explanation which means omnichannel strategies do not impact purchasing behaviour as much as it seems. Higher-value customers are more likely to choose to engage with multiple channels and omnichannel offers. In other terms, a self-selection bias exists which makes it hard to unpack whether omnichannel customers spend more because of using various channels or whether they use them because they were always going to spend more. My colleagues Sean Sands, Carla Ferraro, Jessica Pallant and I investigated this issue in an academic study that will soon be published in the International Journal of Retail & Distribution Management. We found that a large proportion of the higher purchase value of omnichannel shoppers is due to self-selection, but that omnichannel strategies still have value for retailers. To explain these findings in more depth, it is useful to first briefly explain self-selection and how it affects retailers. What is self-selection and why is it a problem for retailers? In research terms, self-selection is a bias that occurs when there is a lack of randomness, and individuals choose themselves whether, or how, to participate, or are not randomly allocated to groups, but  choose their own group or whether to participate in something. This is a common issue in political polling, as individuals with more extreme views may not wish to share those views publicly. This means the data collected is biased as it does not cover all potential views. A similar issue occurs with online consumer reviews. Consumers are most motivated to share a review if they have had a negative experience, which can bias the overall spread of reviews. A similar bias occurs when comparing groups that individuals have chosen to belong to. This makes it difficult to compare the groups accurately, as there may be differences due to why people chose to belong to the group in the first place. This is why scientific experiments usually randomly allocate participants into groups such as those that are testing a new treatment for a medical condition, and those who receive a placebo. This random allocation helps remove any potential bias from individuals choosing whether to receive a treatment or not. A common example of self-selection bias in retail is loyalty programs. When looking at the value of a loyalty program, most retailers will see that loyalty members spend more, and are more loyal than non-members. However, we also know that consumers who were already shopping frequently with a brand are more likely to join the loyalty program as they stand to receive more benefits from it. My wife and I joined the Bonds Kids loyalty program shortly after our son was born, as we knew we now would be shopping for a lot of kids’ clothes. If you compared our shopping behaviour with a non-member, you would see that we spend a lot and are very loyal to the brand. However, that doesn’t mean the loyalty program has made us spend more. We were always going to spend a lot, and so joined the loyalty program (i.e. self-selected) to receive benefits. As a result, the loyalty program has done little to impact our behaviour, despite our apparent high value to the brand.  We are also not alone in this behaviour. A 2007 study led by Jorna Leenheer of the University of Amsterdam showed that the real value of loyalty programs is seven times smaller than it first appears because of this self-selection problem. Self-selection and omnichannel retailing Similar self-selection issues occur when trying to compare the relative purchase value of omnichannel and single-channel consumers. Like loyalty program membership, higher-value consumers are more likely to engage with a brand across multiple channels as they are more interested in the product category and the brand itself. Continuing my personal Bonds Kids example, my wife and I have now followed their social media accounts to keep an eye on new styles and upcoming sales. Once again, this social media interaction isn’t the reason we spend a lot, the fact we spend a lot means we have more to gain by engaging across these additional channels. Similarly, consumers are more likely to search for information across multiple channels when planning a high-value purchase as there is more benefit to them of doing so. For example, many of us are now faced with setting up home offices, and may be considering a new computer or laptop. As a relatively high value and important purchase, we are likely to look for as much information as possible, such as multiple retailers’ websites, online reviews, and maybe even an in-store visit. Again, this doesn’t necessarily mean looking at these channels will lead to a more expensive purchase, rather it is the expensive purchase that led to searching across channels. These are the issues my colleagues and I tried to address in our forthcoming paper.  We did this by focusing specifically on research shoppers, consumers who search for information across multiple channels prior to a purchase. A common example of this is ‘webrooming’ – searching a retailer’s website prior to an in-store purchase. We surveyed close to 600 Australian consumers about a recent purchase of either clothing or consumer electronics. We measured how much consumers spent and the channels they interacted with prior to and during purchase. We also took multiple measures of reasons consumers may have self-selected as research shoppers, such as the amount they anticipated spending, how involved they were in the purchase and product category and so on. We then used a technique known as Propensity Score Matching to compare the purchase value of research shoppers and single-channel shoppers while accounting for self-selection. Our key findings were: Before accounting for self-selection (ie. a simple comparison) research, shoppers spend on average three times more than single-channel shoppers. This is in line with the famous quote from Glen Senk. However, after we account for all the reasons higher-value shoppers may have self-selected as research shoppers, the difference dropped to closer to two times. There are key differences across categories. Self-selection has a much bigger impact in electronics than it does in clothing and fashion. Looking at individual channels, online channels such as a retailer’s own website have the most positive impact on a consumer’s final purchase value after accounting for self-selection. What this means for retailers Overall, our results suggest that self-selection does play a big role in the apparent impact omnichannel strategies have on the amount consumers spend with a retail brand. This is particularly true in categories like electronics. However, even after accounting for the self-selection bias, consumers who interact with multiple channels prior to purchase still spend twice as much as those who only interact with a single channel. For retailers, this means that encouraging consumers to engage with omnichannel offers still has value, just not as much as it first appears. So, when weighing up benefits and costs of “right-channelling” activities, do not rely on simple comparisons of the purchase value of omnichannel and single-channel customers. Remember that many omnichannel customers were likely already high value customers and that is why they interact with your brand across multiple channels. As the retail industry moves further towards true omnichannel offers, it is crucial that retailers invest in these strategies and get them right. If done well, they can create many benefits for brands. However, we should be cautious when claiming how much additional value these strategies create. Higher value customers self-select to engage with omnichannel offers, meaning the value of these strategies, while still crucial, is lower than it first appears.