The World Association of Research Professionals Vondelstraat 172, 1054 GV Amsterdam, The Netherlands Tel: +31-20-664 21 41, Fax: +31-20-664 29 22

September 2001

SUMMARY OF KEY FINDINGS

• As expected, the ratio of online to offline purchasers varied considerably across the vertical markets studied with Clothing, Health, Home and Toys having the largest proportion of Offline Only purchasers; Electronics and Gifts the largest proportion of Online and Offline purchasers; and Auctions, Books, Music, and Video the largest proportion of Online Only purchasers.
• Key Driver analysis of online and offline spending behaviour yielded predictable results with internet involvement (number of sites purchased at, online purchasing tenure, online purchase frequency) and income showing a strong relationship with spending in all vertical markets.
• Segmentation analysis based upon the number of sites visited and the amount spent online yielded six shopper segments including Browsers, Peripheral Shoppers, Educated Shoppers, Less Involved Browsers, Less Involved Shoppers, and Educated Shoppers.

RESEARCH METHODOLOGY

The results described in this paper are based upon data collected via Harris Interactive and Nielsen/NetRating's eCommercePulse^sm survey. This research examines the online shopping and purchasing habits of American adults. The online survey questionnaire is self-administered to a random sample of over 33,000 adults per month using Harris web-based interviewing technologies. Interviews averaged fifteen minutes in length. The completed interviews were weighted to reflect the characteristics of the online population, as determined through Harris' parallel telephone research.

The sampling procedure utilizes Harris Interactive's standard methodology for online surveys. Potential respondents were drawn at random from the Harris Poll Online database of more than seven million individuals, all of who had elected to participate in survey research with Harris. An invitation to participate in the survey was e-mailed to the selected random sample of adults. The study was described as one about online shopping and consumer satisfaction. Through the e-mail invitation, potential participants were directed to the Harris survey web site.

All interviews were conducted using a self-administered, online questionnaire via Harris' web-assisted interviewing software. This interviewing system permits online data entry of interviews by the respondents. Questionnaires are programmed into the system with several checks in place, mathematical checks, consistency checks, skip patterns, question rotation, range checks, and special edit procedures.

RESULTS

ONLINE VS. OFFLINE SPENDING

Our first step was to understand the difference between online and offline spending. We defined four groups of respondents based their purchasing behaviour across all vertical market categories: Non-Purchasers, Online Only Purchasers, Offline Only Purchasers, Online and Offline Purchasers. Group membership was based on variables that represented only purchase behaviour in the last month. As a result some less frequent online and offline purchasers may have fallen into the Non-Purchaser segment.

As expected, the proportion of online vs. offline purchasers varied considerably across vertical market categories (see figure 1).

• Computer Hardware, Computer Software, Home and Travel contained the largest percentage of Non-Purchasers. These are markets where information gathering is often the most popular online activity. Also, larger ticket items are purchased less frequently and more often offline.
• Auctions, Books, Music, and Video had the largest percentage of Online Only Purchasers and the smallest share of Offline Only purchasers, no doubt because an online purchase in these categories is a relatively easy and low risk transaction. These categories represent the most common first online purchase for new online buyers.
• Clothing, Health, Home and Toys had the largest percentage of Offline Only Purchasers. These types of purchases tend to be high touch activities where the fit, weight or colour of the item, often hard to determine online, is important in the selection/purchasing decision.
• Electronics and Gifts had the largest percentage of Online and Offline Purchasers. Though electronics tends to be purchased offline when the purchase is for oneself (another high touch transaction), it is also a popular gift item that easily purchased online.

As shown in table 1, the average spending levels were consistently highest among the Online and Offline Purchaser group. However, the group that accounted for the largest share of spending varied by category. Online Only Purchasers accounted for the largest share of spending in Auctions, Books, Gifts, Health, Music, Travel and Video. Online and Offline Purchasers accounted for the largest share of spending in Clothing, Computer Hardware, Computer Software, Electronics, Home, and Toys. While the share of spending accounted for by Offline Only Purchasers was always less, it was most prominent in the Home and Toys category.

KEY DRIVERS OF ONLINE AND OFFLINE SPENDING

A key driver analysis was used to determine what variables were most strongly related to online and offline spending. We used this approach to gain an understanding of what drives consumers to spend and make purchases online. The analysis was run in each category. Relationships that were significant and relatively strong were identified.

We found it interesting that the drivers of online and offline spending did not vary much by category, though the drivers that were identified by this analysis were not a surprise. For every category, involvement measures (such as number of sites in category purchased at, number of sites in category considered) were strongly related to online and offline spending. Overall internet involvement was related to spending as well. For example, overall purchase frequency and number of categories purchased from are related to both online and offline spending. As expected, income was also consistently related to spending in each category.

Demographics were significant drivers of online and offline spending only for some categories. The number of years purchasing online was a strong driver of online spending in all categories, and was also related to offline spending in a few categories.

• Auctions, Computer Hardware, Computer Software, Home, Music, Toys and Electronics: Income showed a strong positive relationship with online and offline spending in these categories. The number of years purchasing online was also positively related to online spending, with respondents having been online five or more years spending the most online on average.
• Books and Gifts: Education and years purchasing online were positively related to spending in these categories. Respondents having been online for five or more years and the more educated groups of respondents spent more online.
• Clothing: Income showed a strong positive relationship with online and offline spending in the Clothing category. Years purchasing online was positively related to online spending, but negatively related to offline spending. Respondents who had been purchasing online for less than a year spent less online, and more offline, than their more cyber-savvy counterparts.
• Health: Age was significantly related to online spending, where respondents 50 years of age and over spent the most on average perhaps because of increased health needs. Males were found to spend significantly more than females online. Employment status was found to have a significant relationship with online spending, with the self employed spending significantly more than all other groups online. Years purchasing online was positively related to spending, respondents having been online for five or more years spent more online in this category.
• Travel: Employment status was found to have a strong significant relationship with online spending, with the full-time employed and self employed spending significantly more than other groups online. Education and years purchasing online were positively related to spending. Respondent having been online for five or more years and the more educated groups of respondents spent more online.
• Video: Education was found to be strongly related to online spending, with more educated respondents spending more online. Years purchasing online was positively related to spending, respondents having been online for five or more years spent more online in this category. However, years purchasing online was negatively related to offline spending, where respondents who have never made a purchase or made their first purchase within the last four months spent more offline.

SEGMENTATION ANALYSIS

MODES OF SEGMENTATION

Based on the results of the exploration into online and offline spending, we identified multiple measures to be used for segmenting. They included:

• Lifestage,
• Socio-economic Status (Income),
• Purchasing Intensity,
• Net-centricity/Online Involvement/Online Experience.

Because online and offline spending were related to similar variables in each category, we opted for segmentation schemes that were either based on the total sample, or those which could be applied in the same way to each category. The following are some of the modes of segmentation we explored.

1. Lifestage: Lifestage was based on a combination of age, marital status and the presence of children. When profiled, it became clear the age was driving the differences in the lifestage groups and that marital status and presence of children only contributed in limited categories. This approach was not pursued due to the undimensional nature of the segmentation results.
2. Age by Income: We crossed age with income to create nine segments. These segments were very clear descriptively. They differentiated on online and offline spending, but online and offline spending were directionally the same for each segment. The segment that spent the most online (Age 30-49, Income $75K or more), also spent the most offline. These results were not found to be compelling and were not pursued.
3. Cluster Based Solution: To break the market on age, income, online involvement and shopping intensity simultaneously, we used k-means clustering to define segments based on all four variables. We were unable to build measures of online involvement and shopping intensity based on more than one variable. Unfortunately the single variable representations of these two measures were not stable enough to produce reliable cluster solutions. Very few solutions validated, and those that did were not as actionable as the Age by Income segments. We did not pursue this approach any further.
4. Category Involvement Segments: The category involvement segments were defined for each category by crossing the amount spent online with the number of sites visited. The resulting segments varied on category involvement/experience and category purchasing intensity. We found the result of this segmentation to be the most compelling and choose this approach to present in the remainder of this paper.

CATEGORY INVOLVEMENT SEGMENTATION - ELECTRONICS

The category involvement analysis yielded the six segments described below:

• Browsers visited a high number of sites but did not buy online this category. These are shoppers who make a purchase offline after researching it online. Browsers also tend to be newer to the internet but spend a significant amount of time online.
• Peripheral shoppers visited a high number of sites and spent a modest amount online in this category. These shoppers also spend a moderate amount offline and have the highest total number of online purchases of any segment.
• Educated Shoppers visited a high number of sites and spent the greatest amount online in this category. These shoppers tend to spend a great deal of time online.
• Less Involved Browsers visited a lower number of sites and did not buy online this category. These shoppers spend less time and money online overall. They tend to have lower incomes and spend less offline as well.
• Less Involved Shoppers visited a lower number of sites and spent a modest amount online in this category. These shoppers tend to have higher incomes and spend less offline.
• Site Loyal Shopper visited a lower number of sites but spent the greatest amount online in this category. These shoppers tend to have higher incomes and spend a moderate amount offline.

As seen in appendix 1, the category involvement segmentation of the electronics market yielded interesting demographic/cybergraphic definitions of the six shopper types defined above. The most notable differences appear between the Educated Shopper and the Site Loyal Shopper – where both groups show high online spending in electronics but differ in the amount of browsing or surfing they do prior to making a purchase. (See appendix 1.)

The Educated Shopper visits many sites before making a purchase and is more likely to be male (76%) with higher income, more time online, better educated and spend more money online in electronics (average = $141 per purchase). The Site Loyal shopper visits fewer sites before making purchase and is more likely to go to a favourite site to buy electronics. Though the Site Loyal Shopper is still likely to be male (59%), there are many more female Site Loyal Shoppers (41%) than female Educated Shoppers (24%). The Site Loyal Shopper also spends less time online, most likely because they do not spend time shopping at as many sites at the Educated Shopper. Like the Educated Shopper, the Site Loyal Shopper also tends to be better-educated, higher income and spend more money online in electronics (average = $170 per purchase).

Appendix 2 presents the six shopper types web site specific activity in the Online Electronics market. We see that the two groups of non-purchasers, Browsers and Less-Involved Browsers, tend to visit the same sites and in the same rank order of preference. Moreover, BestBuy.com is the primary site visited by these shoppers by a significant margin. BestBuy.com is also the most popular site for all shoppers who visit a high number of web sites (Browsers, Peripheral Shoppers and Educated Shoppers). Peripheral Shoppers and Educated Shoppers differ in the sites where they purchase online with Peripheral Shoppers preferring to purchase at ebay.com, Amazon.com or RadioShack.com while Educated Shoppers prefer to buy at BestBuy.com, Walmart.com, Staples.com and Amazon.com. Finally, we see the Site Loyal Shoppers visiting Amazon.com in much higher proportions than other sites, though they tend to buy at both Amazon.com and Handspring.com. (See appendix 2.)

Figure 2 presents the distribution of purchasers across the top ten web sites in the online electronics market. As discussed previously, Amazon.com and Handspring.com enjoy the largest proportion of Site Loyal Shoppers while ebay.com and RadioShack.com attract the Less Involved Shoppers. A warning sign for their online operations, RadioShack.com and Staples.com have the lowest proportion of Site Loyal Shoppers among the top ten electronics web sites. It also may be their intention to use their online presence primarily to drive offline sales.

Similar trends can be observed in figure 3, which shows the distribution of revenue across the top ten websites in the online electronics market. Handspring.com and Amazon.com recognize the most revenue from Site Loyal Shoppers while Sony.com captures the most revenue overall as well as a significant proportion from Peripheral Shoppers. Also notable is that while Sony leads in total revenue share, Amazon has more purchasers. This is most likely due to Sony selling higher ticket items with minimal discounting while Amazon offers a wider selection of Electronics and provides more items in the lower price ranges. RadioShack.com and OfficeMax.com capture the smallest share of the online electronics revenue. (See figure 3.)

The distribution of spending also varies by type of purchase, that is, whether the purchase was for business, personal or gift purposes. Figure 4 shows the Educated Shoppers spend the most overall followed by Site Loyal Shoppers, Peripheral Shoppers and Less Involved Shoppers. Educated Shoppers also lead in personal spending while Site Loyal Shoppers lead in business and gift spending. When shopping for oneself, surfing to find the best deal may be part of the fun but many business shoppers have specific sites selected because they meet their needs with minimal hassle. For example, certain sites may cater to the business shopper by setting up their business account and providing easy ordering, billing and delivery.

CONCLUSIONS

The internet is an incredible force, which now influences approximately 59% of the adult population in the United States. Now that the hype and fervour surrounding everything 'dot com' has subsided, online shopping remains a critical element in most retailers' channel strategy.

We must, however, recognize that being online does not mean buying online. Over half of the internet users we interviewed did not make a purchase online in the last month. Over one-quarter of visitors did not make a purchase in the categories they considered shopping in. Clearly, the influence of online shopping is not limited to visiting a site and making a purchase. A large portion of online shoppers is making offline purchases after visiting relevant sites online. This has been a problem for 'pure play' online retailers in the past and has limited them to categories most conducive to online purchasing. In fact, many of the remaining 'pureplay' e-tailers are now partnering with their bricks and mortar counterparts to capture the online shopper/offline buyer market (Amazon.com and ToysRUs for example).

Our findings regarding online shopping activity are summarized below:

• The penetration of online shoppers is highest in the music and books categories, as expected.
• The percentage of internet users making offline purchases as a result of online shopping is highest for clothing, electronics and toys. Among those making offline purchases, over 40% have only purchased offline in the past month for these categories.
• Even in books, music and video about 30% of online purchasers also made offline purchases. (Only in the Auctions category, where there are no direct offline alternatives, did Online Only purchasers represent a larger share.)
• People who made purchases Online and Offline did not always represent the largest share of purchasers, but consistently had the highest spending levels.
• For most categories Offline Only shoppers had higher average spending levels than online shoppers did.

Who is spending the most online? The description of high online spenders is no surprise. Levels of online spending are most strongly related to online experience. Respondents who have been online longer and purchasing online longer are spending more online. Income and years of education are also related to levels of online spending in many categories. Levels of offline spending (as a result of online shopping) are not as easily explained. In some categories the offline spending is highest for the least experienced online shopper.

As noted in the discussion on modes of segmentation, finding meaningful segments using the existing eCommercePulse database was not a simple task. The survey was not specifically designed to yield segmented groups and thus considerable exploration was required to identify the most appropriate segmentation scheme.

In the end, we believe the category involvement segmentation yielded an enormous wealth of information about online shoppers in the electronics market. The six shopper types defined by this research include Browsers, Peripheral Shoppers. Educated Shoppers, Less Involved Browsers, Less Involved Shoppers and Site Loyal Shoppers. The unique shopping behaviour of these segments could be used to design and implement more effective marketing programs, which target specific marketing messages for each of the segments. For example, Amazon.com has the largest percentage of Site Loyal Shoppers- those shoppers who tend to visit fewer sites but spend more online. A custom loyalty program designed to encourage and retain these customers would be a prudent move.

We believe that there are exciting opportunities for additional research on this topic including:

• examining the trends in offline shopping habits to determine if offline purchasing as a result of online shopping is a growing trend as well as exploring the seasonality of shopping habits, both online and offline, and how high purchase seasons affect the composition of purchasers;
• investigating the category involvement segments in other categories to better understand how these segments apply to other retail areas;
• exploring the attitudes, needs and values of these segments through additional research to better understand the thought behind their behaviour.

APPENDIX 1: CATEGORY INVOLVEMENT SEGMENTATION: ELECTRONICS

APPENDIX 2: CATEGORY INVOLVEMENT SEGMENTATION AND ONLINE ACTIVITY - ELECTRONICS

http://www.warc.com

ESOMAR. All rights reserved