Archive for April, 2009
“It demonstrates incredible marketing capability.”
The Internet is now a key source of info for the majority of people in the UK. Consequently, search engines are one of the first places people turn to for information on everything from celebrity culture to the global financial crisis. Therefore it can be informative to use Hitwise search data as a way of monitoring not only online behaviour, but also the reaction to events in the ‘real’ world.
We often see these correlations between online and offline data in the financial services world. For example, searches for ‘house prices’ broadly follow real house prices: increasing in the boom times, but falling as the credit crunch began to bite. But what about interest rates: how have people responded the massive drops that have taken place over the last six months? The chart below illustrates two data sets: the pink line represents the Bank of England interest rate over the last 3 years; the blue line is the weekly UK search volume for the term ‘interest rates’.
Up until Autumn 2008 – broadly speaking – searches for term ‘interest rates’ increased in line with actual interest rates. There were obviously weekly fluctuations and noticeable peaks during the weeks when interest rates actually changed, but the correlation is there to see.
However, searches for interest rates increased dramatically last Autumn, as the Bank of England’s monthly decreases became more significant. Searches for ‘interest rates’ reached their highest ever level following the 1.5% drop at the start of November 2008 and since then they have peaked each time that the rates have changed. But the height of those peaks has decreased in line with falling interest rates, re-instating the correlation between searches and actual rates.
It is also interesting to see where people go after searching for interest rates – and how this has changed over time. Looking back to the boom times and picking week when the Bank of England increased interest rates (w/e 11/11/06), 40.7% of people searching for ‘interest rates’ went to a website in our Banks and Financial Institutions category, while 29.6% went to a News and Media site. Fast forward to the last downward adjustment – w/e 07/03/09 – and more people (50.9%) went to a News and Media site than visited the Banks and Financial Institutions category (29.6%).
So, the general increase in ‘interest rate’ searches is as much down to curiosity as people making financial decisions. Presumable this is because interest rate changes are now perceived to have a wider economic significance than simply feeding into mortgage rates. Then again, it may be that news websites have realized that the economic crisis is a good story, and have therefore become more effective at picking up traffic from economic search terms. The truth, I suspect, lies somewhere in between these two explanations.
Back in February, my colleague Heather Hopkins wrote about the shrinking gap in visits between the two leading online maps, Google Maps and MapQuest. As of the week ending April 11, 2009, Google Maps has now edged past MapQuest with a higher market share of visits. While the share of visits is higher for Google Maps, the average visit time on MapQuest remained higher at 10 minutes and 51 seconds as compared to 7 minutes and 24 seconds on Google Maps. Google is the top referral sources for both websites, but is considerably higher for Google Maps with 61% last week while the share of visits referred to MapQuest was 25%.
The case still remains that MapQuest is a strong brand and is the leading search term driving traffic to the Maps category, but there appears to be some erosion that has taken place over time. The share of traffic from the search term ‘mapquest’ (which represents 32% of the clicks driving traffic to MapQuest) to the Maps category has declined 32% for the 4 weeks ending April 11, 2009 when compared to the same time frame one year ago.
Albeit amidst some controversy following the launch in the UK, the Google Maps feature ‘Street View’ has been very popular and the share of searches on ‘street view’ increased 29% for the 4 weeks ending April 11, 2009 when compared to the same timeframe one year ago. Not surprisingly, Google Maps is the main beneficiary of the searches and received 74% of the traffic from the search term during the same timeframe.
With a tight race between Google Maps and MapQuest, will innovations such as street views be the deciding factor?
While perusing this week’s fast moving search terms (i.e. the search terms in each category that experienced the biggest increase in volume between last week and the week before), the list for our Shopping and Classifieds category caught my eye. Last week the three fastest moving search terms for retailers were ‘tesco opening times’, ‘tesco opening hours’ and ‘asda opening times’; while ‘sainsburys opening times’ and ‘sainsburys store locator’ also featured in the top 10.
I was curious to see whether these searches had increase year-year on year, and how they compared with Christmas time, so I created the chart below. As you can see, search volumes for a range of supermarket-related store opening times terms were higher during Christmas, but increased this Easter when compared with last year. As the biggest of the three retailers, Tesco receives the most searches – followed by Asda and Sainsburys
Looking at the search term variations reports for the terms ‘opening time’, ‘opening hours’ and ‘store locator’, all are popular queries. Last week (w/e 11/04/09), UK Internet users searched for 12,100 distinct terms containing the phrase ‘opening times’, 7,200 containing ‘opening hours’ and 500 containing ‘store locator’. As with searches containing the word ‘sale’, the most popular terms are those including shop names. The supermarkets top all three lists, while other popular searches include ‘post office opening times’, ‘b&q opening times’ and ‘b&q store locator’.
Generic – i.e. not shop-specific – ‘opening times’ related searches are less popular, but still significant. ‘good friday opening times’ was the seventh most popular search term to contain the phrase ‘opening times’ last week, while terms such as ‘tesco easter opening times’ and ‘asda easter opening times’ were more popular than generic terms such as ‘easter sunday opening times’ and ‘easter opening times’. It was a similar situation for variations on ‘opening hours’.
Although there were more shop-specific ‘opening times’ related searches over the Christmas period, there were more generic searches in the run up to Easter. As the chart below illustrates, there were four times as many searches for ‘good friday opening times’ last week as there were for ‘boxing day open times’ during Christmas week.
The generic ‘opening times’ related searches also represent missed opportunity for retailers and affiliates. Currently just 64.7% of ‘good Friday opening times’ searches are successful, meaning that a third of people searching for the term last week didn’t click on a link from the search engine results page (SERP). Further evidence of a content gap / SEO opportunity is provided by sites that did manage to pick up traffic from the generic terms. The top to recipient of traffic from ‘good Friday opening times’ searches last week was The AnswerBank, while Yahoo! Answers was top for the term ‘easter opening times’.
While traditional advertising media is losing its effectiveness, new opportunities, such as video games, are presenting themselves to advertisers.
Back in January, Circuit City announced ceasing all operations, so we thought it would be interesting to analyze which retailers have benefited from their exit in the market. The brand Circuit City is very well known and drove a significant share of searches from consumers using the name to navigate. As a result of the website and store closings, the share of visits that are going to the Circuit City website that included a search term from a portfolio of Circuit City’s branded terms have declined from 87% during the week of Christmas to 68% for the week ending April 4, 2009. Among Circuit City’s former competitors in the electronics arena, the main beneficiaries of the Circuit City searches are Radioshack.com and Bestbuy.com.
Both of the retailers are receiving the search traffic from the portfolio of branded terms in opposite ways. For Radioshack.com, 100% of the traffic from the Circuit City Terms was from paid search over the 4 weeks ending April 4, 2009. With Bestbuy.com, all of the search traffic was driven by organic search results during the same timeframe.
While Circuitcity.com continues to receive visits as the results of searches and other sources, the only content on the website is the message about the store closings. Once they reach that message, visitors often move on to other retailers to look for products. Here there are also two retailers who are the main beneficiaries of the traffic from Circuitcity.com, Bestbuy.com and Walmart.com. During the week of April 4, 2009, Bestbuy.com captured 20% of the downstream traffic from CircuitCity.com and Walmart.com followed with close to 10%. Kmart.com and Sears.com each captured 4% and 2%, respectively.
This week kicked off with the Associated Press voicing concerns about how websites obtain permission to distribute content and share revenues. A debate has sparked about the ‘fair use’ doctrine (publishing a headline and/or sentence from story) and tracking the legal use of the content. Search engines and news aggregators were all somewhat implicated as those who may be misusing content, so we decided to take a look at their impact upon driving traffic to news websites, which ultimately generate ad revenues.
Like many categories, search is one of the main sources that drive traffic to the News & Media category, referring nearly 22% of visits in March 2009. Branded searches for news properties represent a large share of the top search terms driving traffic to the category. Another major source is the front pages of portals such as Yahoo! and MSN, including the personalized versions like My Yahoo and My MSN. In comparison, social networking websites and blogs referred a far smaller share of visits to News & Media websites from links or references posted on their pages.
Traffic moving between News & Media websites is also a main source of traffic equal to search at 21.63% in March 2009. However, the shares of the sources are distributed broadly across a wide number of websites where the highest share coming from any one website was only 2.28%.
Although several of the online aggregators are at the heart of the content distribution argument, they do successfully send visits to news properties rather than keeping them upon their own websites. Two sources that have increased the share of referrals in March 2009 when compared to March 2008 are Drudge Report and The Huffington Post. Others that have remained consistent include Google News, Yahoo! News and My Yahoo.
To piggy back on a post my colleague across the pond, Robin Goad, did yesterday, we thought we’d revisit some of the stats to better understand how things would look if Google were to acquire Twitter.
As the chart below illustrates, US Internet visits to Twitter have increased 194% since the start of the year and nearly 7000 times in the last 12 months (March 08-09). Last week (w/e 04/04/09), Twitter was the 69th most visited website in the US, up from 74 the week before, 86 two weeks ago and 105 three weeks ago. It was the 8th most popular social networking site. The website has enjoyed steady and constant growth. Share of US Internet visits to Twitter have increased every week since mid-November 2008.
I should also add the usual caveat: the service is probably even more popular than our numbers imply, as we are only measuring traffic to the main Twitter website. If the people accessing their Twitter accounts via mobile phones and third party applications (such as Twitterrific, Twitterfeed and Tweetdeck) were included, the numbers would be even higher.
But where does Twitter send its traffic to? It may be the 69th most visited website in the US, but Twitter was also the 66th biggest source of traffic to other websites in the US during March. A few weeks ago we posted data showing where US consumers go after visiting Twitter.
Here is an updated version of the chart from my last post. This time I am using March data and I changed the chart type to ease comparison.
As I mentioned in my last post, “it appears that Twitter is being used as a social network and means of distributing content. This is by no means the only way it is being used – just one standout trend. Twitter.com’s clickstream profile is much closer to a social network than to Search Engines or Email Services. Twitter’s clickstream differs markedly from search engines in that relatively little traffic goes to retail websites and Education (i.e. Wikipedia). It is also different from Email in that less traffic goes to Dating websites and again, to retail and Business and Finance websites.”
In other words, Twitter sends far less traffic to industries that are the big spenders on search engine marketing than does Google.
The sun may be shining, but it’s not quite time for the music festivals yet (it has to start raining for that to happen). However, following the release of the Reading and Leeds Festival line-ups in the NME last week, it is certainly festival season online. As the chart below illustrates, the number of search terms containing the word ‘festival’ has been increasing since Christmas, and notably spiked last week.
Having the exclusive on the line-up announcement was a coup for the NME, and it paid off in terms of traffic. UK Internet visits to the website increased by 37% last week and, as the chart below illustrates, it was the music magazine’s busiest week since last summer. In addition, ‘reading festival’ was the second biggest search term sending traffic to the NME last week after the navigational term ‘nme’. In total, six of the top 10 search terms sending traffic to the site were related to either the Leeds or Reading Festivals.
Looking at the sites visited after the NME last week, 36.7% of downstream traffic went to retail websites. Play.com, which has recently started selling tickets on its site, was the biggest beneficiary. The online entertainment retailer was the number one site visited after the NME last week, picking up 22.3% of the music magazine’s downstream traffic. The NME was also Play.com’s sixth biggest source of traffic last week, accounting for 1 in every 35 visits.
The NME also sent a significant amount of traffic to pure-play ticket resellers, which accounted for 9.5% of the site’s downstream traffic. Over half of those visits went to See Tickets, which was the second largest recipient of traffic from the NME last week after Play.com.
The Reading, Leeds and T4 On The Beach festivals were also the three most searched for music events on ticketing websites last week. This is first time that festivals have taken all of the top 3 positions since we started compiling the data for Music Week last year. For a full list of the top 20 music event (i.e. including festivals and single artist tours) searched on both primary (e.g. Ticketmaster) and secondary (e.g. Seatwave) ticketing websites, make sure to pick up Music Week magazine every Thursday!
Last week I presented on a panel with Professor Arturo Perez-Reyes (UC Berekeley) at an event in San Francisco we put on with Jupiter Research. One story the Professor shared about Toyota was something I hadn't heard before…
He said that Toyota followed the baby boomer generation as a market. They looked at the demographics and spending power of that generation. I think it went something like this… They started with the Corolla, then Celica for when they got into college, then Corona/Camry, then launched Lexus when they had discretionary income. They followed the 'bulge' of spending the baby boomers had.
It's an interesting way to think about the markets you're going after. Is it big? how will it evolve? How will it effect your product strategy?