Last week, AdAge ran a story, Why Search May Not Click for Retailers.   Reactions to the story, which is based on a new study from Nielsen’s online division, have ranged from head-shaking to a quest for a deeper understanding.  To help retailers better understand these ideas, the Shop.org Blog facilitated a conversation between the author of the Nielsen report, Ken Cassar, and George Michie, CEO and co-founder of The Rimm-Kaufman Group, a Chartlottesville, VA-based search agency.  Ken and George will be monitoring this blog post and will be happy to answer any additional questions in the replies.

George Michie, CEO and Co-Founder, Rimm-Kaufman Group

George: The Nielsen data presented by Ad Age suggests that the average retailer gets less than 10% of its traffic from search, and that much of that traffic is navigational rather than what we refer to as “competitive search.”  In our experience that number varies wildly by the type of company.  Online pure plays may generate 50 – 75% of their traffic from competitive non-brand paid search, catalogers more often 15 – 20% and chain store retailers much lower at 5 – 10%.  We see the number vary with the age of the business as well.  Newer companies with less of a loyal following generate a larger fraction of their sales from paid search.  Are our findings inconsistent with yours, or did the Ad Age article misstate your findings?

Ken: We have been tracking how the top 200 online retailers attract visitors to their sites for a few years now, and we have consistently found that, among those sites, IN AGGREGATE, search driven visits account for less than 15 percent of visits to retail Web sites (the number varies from quarter to quarter, but falls in a 10 – 15 percent range).  However, there are a few really important qualifiers that cannot be glossed over:

Ken Cassar, VP, Industry Insights, Nielsen's Online Division

1. We are looking only at the top 200 retail Web sites.
2. We use a rigid definition of a search referral where there needs to have been a click from a natural or paid search link.   If a person searches and then types a URL, that would be assigned to our ‘direct-to-site’ bucket.
3. The methodology that we use to look at the aggregate view of the top 200 is one that assigns a heavy weight to the bigger retailers.  We look at total referrals from each source divided by total visits.  As a result, the bigger retailers such as Amazon, eBay and Wal-Mart exert a high degree of influence over the total number.  When we look at individual retailers, we see a wide variation, even among retailers in the same category.

Obviously, the reason that we do this is because of the clients that we do the work for.

The differences in the data for individual retailers drives very different advice for different types of retailers.  Walmart.com’s marketing strategy should be a lot different than Celticstore.com.  However, I do think that there are important lessons for niche retailers to be taken from what these bigger retailers have learned.  There is a reason that people type in the URLs of certain sites, and engage in navigational searches to certain brands and sites:  They are aware of the brand, it is top of mind when they think about what those retailers sell and that increases the likelihood of a direct visit, rather than one mediated by search engines.

Don’t get me wrong:  For the vast majority of retailers if search isn’t the most important part of their online marketing strategy, it’s next on the list.  The key point that I make, though, is that brand matters.  While the benefits of brand development are less quantifiable, they are of critical importance for all retailers.  Too many retailers that I see don’t pay enough attention to brand dividends because they are difficult to track.  The fact that the majority of visits to large retailers are the result of consumers typing in those retailers’ URLs or searching for their brands is the easiest proof that I can offer.

George: Does Nielsen suggest that money currently going into paid search should be spent elsewhere, or is the point more that other off-line media still drives the lion’s share of business for most non-pure play retailers?

Ken: I don’t think that there is one answer to this question for all retailers.  If I had to generalize, I would say that retailers allocate too few dollars to online marketing.  Multi-channel retailers only track the online sales benefit of the marketing that is trackable (like search) and ignore the offline sales benefit.  Most retailers, I think, under-invest in brand oriented advertising, whether it is display, video, sponsorship or in social media environments.  One of our biggest (and most fun) challenges at the Nielsen Company is helping retailers understand these difficult-to-measure offline effects.  I would like to see niche-oriented online retailers devoting more brainpower to creative ways to create exposure to their brands.  SEO and SEM are going to represent the lion’s share of their acquired sales, but they’ll work better if consumers have a level of awareness of their brands.

George: Paid search is often the largest ad spend among online channels but is usually small in comparison to what chains stores spend on TV and circulars.  Do you suggest that more money spent on those channels will produce better return on investment?

Ken: When we do analysis for advertisers looking at the efficacy of advertising in different media (traditional and online) I continue to be amazed at the strong performance of television advertising on an efficiency basis.  Despite the fact that it often constitutes the majority of an advertiser’s budget, it tends to drive at least an equivalent share of ad driven sales.  The decline in newspaper circulation does create challenges for retailers.  Over the past 20 years, newspaper circulation has dropped by about 20 percent (that number is probably a bit conservative).  The result is that printed circulars do not have the universal reach that they once did.  The biggest opportunity for retailers that were historically big print circular advertisers is figuring out how to disseminate their promotions online, rather than just putting their circulars on their sites.  Both search and display are well positioned to play this role.  The good news is that in our research, online ROIs  (measured against offline sales) have been quite strong.

George: We certainly agree that navigational search on a retailer’s trademarks shouldn’t be credited to paid search, but if someone searches for “Big Screen TV” do you think electronics retailers should credit that ad for the sale, or a TV ad they might have been running featuring their TV selection?

Ken: Attribution of a sale to a marketing event is clearly one of the biggest challenges that marketers have.  There are steps that can be taken, using well constructed test/control studies or regression based analyses to mitigate compound advertising effects, but it is not easy, and it is never perfect.

George: Some people interpreted the Ad Age article to suggest that Nielsen thinks paid search is bad and display advertising is good.  Is that your point?

Ken: Absolutely not.  Search is great.  If anything, retailers (particularly multi-channel retailers) should be doing a lot more with search than they are.  However, brand matters, and the devices, including display advertising, that allow retailers to develop brand are a lot more important than many people think.

George: We generally argue that paid search is not a demand generating channel, it is a demand capture channel and that there will always be a need for the demand generators (print ads, TV, display ads, etc.)

Ken: I like that distinction a lot.  While display and search compete for dollars, they are apples and oranges.

George: Have you folks studied the data on display ads carefully?  Many carefully controlled tests (where public service announcements are shown a fraction of the time) fail to show display ads producing any lift in site traffic.

Ken: I’ve got a few thoughts on that…

First let me clarify the data.  Our data is NOT showing that display ads generate more traffic than search.  In fact, we haven’t quantified the share of visits driven by search.  We quantified search driven traffic, direct-to-site traffic, and comparison shopping site driven traffic.  The remainder we classify as third party/other.  It includes clicks on display ads, as well as affiliate driven traffic, and email.   This is probably the biggest misunderstanding of the discussion that has followed the Ad Age article.

However, I do believe that the tools used to commonly quantify the effect of display advertising look at timeframes that are too narrow.  And that a lot of the display advertising used by advertisers has been too focused on trying to drive an immediate transaction.   Here’s an example from a different vertical that, I think, will shed some light on the issue.  We’ve run 300+ studies for consumer packaged goods manufacturers quantifying the impact of online advertising on offline sales.   When I run a simple correlation of the relationship between click through and ROI the correlation co-efficient is…drum roll….negative .07.   The takeaway from this is, I think, that to really understand the impact of display advertising, you need to look at it through a wide angle lens.  The low hanging metrics commonly used to measure ad impact aren’t great at that.

George: Would you suggest that retailers assess their own data, or rely on industry averages in making marketing decisions?

Ken: I would recommend that retailers take a hard look at their own data, and challenge themselves to measure the effects that are hard to see.  Secondarily, I think that there is a lot of benefit to come from looking at how other companies are driving traffic.  The right answer lies between the two.

George: What should be the big take away from the Nielsen research on this topic?

Ken: Brand matters.  While this might be more evident when we look at big retailers, it is an important takeaway for all retailers.

Editor’s note: George followed up this Q&A with additional thoughts on the Rimm-Kaufman Group blog.