1 White Paper SimpleRelevance Research Report Fail: An In-Depth Evaluation of Top 20 Internet Retailers Personalization Capabilities ABSTRACT In today s retail world, personalization is key. From ed grocery store coupons to smartphone apps, consumers today expect retail messages to be tailored to their individual interests. In a 2014 survey conducted by Forrester, 70 percent of executives interviewed believed that personalization was of utmost importance to their business strategy. However, only 17 percent of marketing leaders are going beyond basic transactional data to deliver personalized messages to consumers. In July 2014, SimpleRelevance analyzed the campaigns of 20 top internet retailers within the 2013 Internet Retailer Top 500 to determine how effectively leading companies were leveraging their datadriven personalization capabilities. The results were shocking: household names such as Macy s, Target, and The Home Depot consistently failed to execute basic forms of personalization. This report is an in-depth review of the parameters and findings of our research.
2 2 Contents, Sophia and Chet CONTENTS Abstract Sophia and Chet The Macy s Example Internet Retailer Landscape Internet Retailer Rankings Revenue vs. Personalization Component Analysis Product Recommendations Type of Recommendation Image Analysis Time of Day Optimization Frequency Optimization Subject Line Other Findings Final Verdict About SimpleRelevance SOPHIA AND CHET Our team created two fictional customer personas to illustrate how retailers leverage customer information, transaction history and interaction data to offer customers a personalized shopping experience. Name: Sophia Wright Gender: Female Age: 34 Name: Chet Hammland Gender: Male Age: 34 Sofia and Chet were assigned Gmail accounts, allocated unique sets of user attributes and behaviors, and led to make purchases from 20 different retailers. Our team analyzed the subsequent s that they received to see how well each retailer was speaking to Sophia and Chet based on their unique attributes. Using Sophia and Chet s accounts, we purchased one item valued between $20-$40 from each company, signing up for newsletters during the checkout processes. Once initial purchases were made, half of the items were randomly selected, and returned. Over a six-week period, we tracked and analyzed specific elements of personalization: - Product Recommendations - Image Optimization - Time of Send - Send Frequency - Subject Line We were surprised to find that many of the largest retailers were not personalizing these attributes, and as a result were not achieving their full sales potential.
3 3 The Macy s Example THE MACY S EXAMPLE One example of a retailer that failed to personalize their s effectively was Macy s. Sophia and Chet both purchased gender-specific items from Macy s online store and signed up to receive Macy s newsletters. Here is an example of a newsletter that both Sophia and Chet received: Sophia s purchase: Ralph Lauren Women s Belt - $33.18 Chet s purchase: Men s Polo Shirt - $42.31 This contained eight different categories of product offerings, five of them directed at females, two directed at both genders, and one directed at males. While it made sense for Sophia to receive these product recommendations, it was evident that Macy s did not leverage the data provided by Chet s purchase or his profile to send him personalized content. Macy s knew Chet s gender based on the information given during the check out process, as well as based on his recent purchase of a men s Polo shirt. Macy s could have leveraged this information about Chet to send him more targeted product recommendations, rather than a generic newsletter more suited to female customers. However, Macy s is the not the only company that struggles with personalization. According to a joint study by Econsultancy and Adobe, 63% of com-
4 4 The Macy s Example, Internet Retailer Landscape panies surveyed do not have the ability to target personalized web content in real time, making content personalization impossible. Many of the other companies whose campaigns we analyzed struggled with similar issues. In fact, of the 418 total s we received, only 37 offered personalized product recommendations. 78 of the s received suggested that their product recommendations were tailored for the customer, with messages such as Recommended for you or Items just for you. However, over half of those supposedly personalized s were in fact composed of generic content. INTERNET RETAILER LANDSCAPE The online retail marketplace in the United States is one of the fastest-growing sales channels in the country, with more than 100 million Americans making online purchases each year. E-commerce alone accounted for 7.31 percent of all retail sales in 2012, and has grown by 287 percent in the last 10 years. By 2017, revenue generated through e-commerce sales is expected to increase at an average annual rate of 10.5 percent, resulting in $370 billion of potential earnings that year. This swift growth is too groundbreaking to ignore. As e-commerce sales continue to increase, thousands of companies across the United States will battle to earn valuable market share. Internet Retailer Rankings To provide insight on this competitive market, trade publication Internet Retailer ranks companies according to the revenue they generate through e- commerce sales. We initially selected 20 companies from the 2013 Internet Retailer Top 500 to use for our research, 13 of which were in the top 20. Of the 20 companies analyzed, Walgreens, Sears, and CVS failed to send a single promotional . To keep the data reliable, we removed them from our data set, leaving us with 17 total companies to evaluate. These 17 companies ranged in industry from apparel to home appliances and electronics. All companies chosen, with the exceptions of Amazon and Netflix, have brick-and-mortar locations across North America, but also rely heavily on sales generated online. For example, in 2013, 45 percent of Staples total revenue and 7 percent of Best Buy s $42 billion in revenue resulted from online sales. Forecast: US Online Retail Sales, 2012 to 2017 $231 $262 $291 $319 $345 $370 (US$ billions) Year-on-year growth Share of total retail sales 14% 13% 11% 10% 8% 7% 8% 8% 9% 10% 10% 10% Source: Forrester Research, Inc. Online Retail Forecast, 2012 to 2017 (US).
5 5 Internet Retailer Landscape Revenue vs. Personalization In order to evaluate the quality of personalization present in retailers campaigns, SimpleRelevance applied a proprietary formula that takes components such as time of send and personalized product recommendations into consideration, which we evaluate in-depth later. After applying this analysis to the s of the 17 retailers evaluated, we re-ranked the retailers based on completeness of personalization. Below is a side-by-side comparison between the retailer rankings by revenue and by quality of personalized s. Ranking Based on Revenue Ranking Based on Personalization Internet Retailer Rank Company Revenue (in billions) Our Rank Company Change in Ranking 1 Amazon $ CDW 8 2 Staples $ Amazon -1 3 Apple $ Wal-Mart 1 4 Wal-Mart $7.7 4 Dell Sears $4.2 Office Depot $4.06 Dell $3.9 Netflix $3.61 BestBuy $3.35 Macy s $ Lowes 11 Men s Wearhouse 14 Home Shopping Network 5 BestBuy 1 Target 4 12 CDW $ W.W. Grainger 2 15 W.W. Grainger $ Home Depot Target $1.93 Home Shopping Network $1.45 Walgreens $0.90 Lowes $0.76 Home Depot $0.74 CVS $0.28 The Children s Place 294 Men s Wearhouse $0.22 $ The Children s Place 7 Macy s -2 Staples -12 Netflix -7 Apple -13 Office Depot -11 Sears N/A Walgreens CVS N/A N/A
6 6 Component Analysis COMPONENT ANALYSIS At SimpleRelevance, we help clients leverage their personalization capabilities by optimizing several key components. In our analysis of retail companies newsletters, we took the following components into consideration: - Product Recommendations - Image Optimization - Time of Send - Send Frequency - Subject Line Product Recommendations One of the first and most important aspects of personalization is tailored product recommendations. While it may seem obvious that retailers should include personalized product recommendations in marketing s, many of the evaluated companies failed to meet this simple criterion. Out of the 418 s we received during the evaluation process, 104 did not include any product recommendations at all. 16 of these 104 s did contain product feedback requests, asking customers to give feedback (in the form of a rating or survey) on products that they ordered. These retailers failed to capitalize on a unique opportunity: product feedback s are one of the best opportunities to make new product recommendations. The best time to ask for more is after a yes. Type of Recommendation Another category of components that we analyzed was whether product recommendations were completely individualized, or recommended by category (i.e. 13 MacBook Pro with Retina Display in lieu of simply Laptops ). We observed that 46 percent of recommendations were generated on an individual basis, and 54 percent were delivered based on categories. While this research report does not favor one method of determining recommendations over the other, it does maintain that retailers should work with their technology partners to create program goals up front. Once a program with clear parameters is established, customer behaviors such as clicks and purchases can be monitored and analyzed to determine which recommendation model best suits a company s specific business needs. Image Analysis In an increasingly visual world, the presence of personalized images in an campaign can also have a marked impact on how customers interact with content. Although most retailers include one or more images in their s, many lack image personalization. Our data showed that only 10 percent of s contained personalized images. 19 s failed to include any images at all. 10% Because the customer has just made a purchase, s(he) is in a state where s(he) is likely to purchase again. By sending product recommendations at this time, retailers can capitalize on the customer s recent yes if s(he) had been satisfied with the last purchase, and even turn a no into a yes if the last purchase did not meet expectations. 90% s containing personalized images s without personalized images
7 7 Component Analysis A noteworthy success from our research was Amazon s recommendation following the purchase of an indoor basketball hoop. The following included product recommendations specifically tailored to indoor hoops; a very relevant option based on the purchase our user made. Another important component of successful personalization is time of day optimization. Not all users check their inboxes at the same time of day. In fact, while they often check their s several times during a day, 60 percent of users only click through s during a single hour each day. We call this the Magic Hour. According to our research, this Magic Hour varies by each individual user. Because an optimal time period exists, marketers have a great opportunity to leverage their customer data and technological capabilities to optimize send times across their full subscriber lists. In the case of our users Sophia and Chet, we set the Magic Hour to be the period between 9-10am. If an was received outside of our users preferred time period, the would be pushed to the bottom of the customer s inbox, causing the click-through potential to decrease drastically. Time of Day Optimization Four of the total 17 companies we evaluated sent s to both Sophia and Chet within the 9-10am period and resulted in increased click-through rates. Dell, Target, and Men s Wearhouse in particular sent more s within the Magic Hour than any of the other companies analyzed. Astonishingly though, over half of the companies evaluated not only failed to send s during the Magic Hour, but also sent most of their s late in the afternoon and evening. Roughly 35 percent of these s contained daily deals and savings for the day of the send, and were rendered ineffective if opened the following day. Open Time Number of s Opened Time of Day
8 8 Component Analysis s Sent Within the Magic Hour Number of s Sent Company Frequency Optimization In addition to sending s at the correct time of day, retailers can also optimize how often they should send s. According to market research and consulting firm Chadwick Martin Bailey, approximately 70 percent of U.S. users unsubscribe from a business or marketing campaign after receiving too many s. On the other hand, sending too few s will also result in companies missing out on potential online revenue. By monitoring user interaction and adjusting frequencies to match user activity, retailers can find an effective middle ground. In order to determine whether our test companies were adjusting send frequency according to user interaction frequency, we created two more users in addition to Sophia and Chet. Rhonda Wiggins and Bob Dunnes were created as inactive users, signing up for newsletters but not engaging with the retailer s shop or content otherwise. We wished to monitor whether retailers would adjust their send frequencies based on how active our users were. Name: Rhonda Wiggins Gender: Female Age: 34 Name: Bob Dunnes Gender: Male Age: 34 After analyzing the data we collected on our 17 retailers and comparing the information received from both the active and inactive user accounts, we determined that none of the 17 companies varied send frequency based on user interaction. Companies like Home Shopping Network and Men s Wearhouse sent an average of 11.8 and s per week, respectively. Neither of these companies personalized send frequency based on user interactions, as Rhonda and Bob received the same number of s as Sophia and Chet, despite their lack of engagement with the companies products. Surprisingly, 40 percent of the evaluated companies sent less than one per week or didn t send marketing s at all. Over the six-week period, Office Depot set on average 0.3 s total.
9 9 Component Analysis s Sent Per Week Number of s Sent Company Subject Line A strong subject line is imperative to the success of an campaign. It is the first thing that a recipient sees, and will determine whether he or she opens the . Good subject lines are clear, to the point, and attention-grabbing. s with long subject lines that are cut off on computers or mobile devices likely end up in the trash. more than 50 characters. The Children s Place and Macy s had two of the longest subject lines out of all of the companies, at 123 and 77 characters respectively. 10 other companies also had subject lines containing between characters. The below from The Children s Place is an example of a wordy subject line that was cut off due to its long length. Looking at the subject lines of the 418 s we received, we noticed that 109 of them contained Cut-off subject line.
10 10 Component Analysis Longest Subject Lines Number of Characters Company In addition to subject line lengths, we also took into consideration the quality of subject line content. SimpleRelevance research shows that customizing the subject line to include the recipient s name, deals tailored to the recipient, and/or a unique topic of interest can result in a 12 percent increase in open rates. lines, suggesting that a majority of s contained content personalized for their recipients. However, evidence suggests that although companies attempt to entice consumers with the promise of special deals and savings tailored for them, they fail to provide personalized product recommendation within the actual itself. Addressing the recipient of an by name is one of the easiest ways to establish rapport, and 96 percent of the s we received from our chosen retailers did not include recipient names in their subject lines. The only two companies to send more than one with a personalized subject line were Amazon and Wal-Mart. In fact, 70 percent percent of Amazon s s included the recipient s name in the subject line. Unsurprisingly, Amazon generates the most online revenue compared to others evaluated. To the right is a table of the top 10 most frequent words that subject lines contained. The table shows the word your as the most frequently used word by marketers in subject Word Your Daily Extra Special Deal Today s Save You New Online Number of Occurrences
11 11 Other Findings, Final Verdict, About SimpleRelevance OTHER FINDINGS Throughout the course of the project, we came across additional findings that retailers could use to refine their marketing strategies: To understand different ways companies implement marketing strategies, we returned half of the products we ordered and see if it triggered a change. We found that none of the 10 companies who had products returned followed up with an abandonment campaign or suggested a different product. While collecting s, we discovered that certain words within the subject line and body triggered inbox spam filters. For instance, s containing words like help, or subject lines written in all capital letters, would often be flagged as spam and go unnoticed for several days. During the initial account setup process, we arranged for each fictional user s birthday to fall within the six-week duration of our experiment. By the end of the retrieval period, not a single company acknowledged the birthdays. No personalized coupons, special sales, or product recommendations were received, proving these companies missed a huge opportunity to solicit brand loyalty. FINAL VERDICT We at SimpleRelevance believe in the existence of an ideal . How this ideal looks for your business depends on who your customers are on an individual level. Many companies have a wealth of user data at their disposal, but are not leveraging their knowledge at full capacity. Our research showed that many top companies fall short of personalization and instead target the masses through bulk recommendations and generic messages. In short there is a massive gap in the level of personalization that customers demand and retailers ability to provide it. This actually proves there is great potential for growth within the online retail market that can be captured through the deliverance of smarter messages. ABOUT SIMPLERELEVANCE SimpleRelevance is a leading provider of personalized marketing communication, leveraging customers unique geographic, social, and demographic data with individual interests and past purchase behaviors to automatically send personalized newsletters. SimpleRelevance offers marketing solutions to both small and medium sized businesses and larger enterprises across a variety of industries. To learn more or sign up, visit
12 12 WORKS CITED Andrzejewska, Hanna. Best Time To Send [INFOGRAPHIC]. GetResponse. N.p., 9 Oct Web. 17 July Banjo, Shelly. Apple Jumps to Second Place in Online Retail - Corporate Intelligence - WSJ. The Wall Street Journal. N.p., 6 May Web. 17 July Gesenhues, Amy. Study: Personalized s Deliver 6X Higher Transaction Rates, But 70 percent Of Brands Fail To Use Them. Marketing Land. N.p., 6 Feb Web. 17 July Jones, Phil M. Helping Your Business Reach New Heights. Beltone Annual Sales Meeting. Illinois, Glenview. July Lecture. Online Consumers Fed Up with Irrelevant Content on Favorite Websites, Accord- ing to Janrain Study - Janrain. Janrain. N.p., Autumn Web. 17 July Online Retail in the US: Market Research Reports. IBISWorld. N.p., Web. 17 July For more information, visit simplerelevance.com SimpleRelevance 130 East Randolph Street Suite 710 Chicago IL Personalization - The Secret to Better Customer Experience. Forbes. N.p., Web. 15 August Copyright 2014 SimpleRelevance. All rights reserved.