Don’t Be Overwhelmed By Data

All kinds of data exists about how your site is doing in search engines and about your potential customers. So much data, in fact, that the thought of making sense of it all can seem a bit daunting.

Web analytics data, keyword research, search trends, search engine tools for webmasters: where do you start?

People often ask me what data they should be paying attention to, but the answer really depends on your situation: What are your goals? What stage is your site in? From that mountain of data, you may only need a few pieces of actionable information to help you measure your current efforts and make adjustments for the future.

A mature site embarking a new TV-based branding campaign will use very different measurements than a new site relying on grass roots efforts to get the word out.

Below are just a few examples of how you might pick just a few pieces of key data to help measure success.

New Site Launch

What should you measure?

Keep track of how the search engine robots are crawling your pages (from your server logs) and how many of those crawled pages are getting indexed.

How is the data actionable?

Often, site owners measure new site progress by watching to see when the site starts ranking well for particular queries or when the site starts getting traffic from search engines. But before either of those things can happen, the site has to be crawled and then indexed. By watching these two milestones, you can pinpoint any problems early.

If the site isn't getting crawled, make sure that you're not inadvertently blocking pages via the Robots Exclusion Protocol and that search engines know the pages exist (either via an XML Sitemap or incoming links). Monitor the percentage of pages from the total crawled that are getting indexed. If it's a low percentage, make sure that you don't have lot of duplicate content issues (such as from dynamically generated pages). Check the Microsoft Live Search Webmaster Tools Crawl Issues reports to generate a list of URLs that are blocked or that search engine bots are having trouble crawling for other reasons.

What next?

Once pages are regularly crawled and indexed, then you can start looking at how well you're ranking for relevant queries, how much traffic you're getting from search engines, and how qualified that traffic is.

Grass-roots awareness effort

What should you measure?

You want to know which audiences are interested in spreading the word about you and which are interesting in becoming customers. When you begin viral marketing, PR, or other campaign, you likely let people know by sending out press releases, talking to the press, sharing the information on social media sites, or providing details to bloggers who write about your subject area.

Make a list of all the ways you spread awareness, organized by category. Then track the links that are generated during the course of the campaign using a reporting tool such as the one found in the Microsoft Live Search webmaster tools. You should also look at your analytics data to make note of the traffic from each link (as well as details such as the conversions and bounce rate from each link).

How is the data actionable?

It's interesting to track how many links a viral campaign produced, but it's much more useful to know which of your efforts sparked those links. A technology site may find that getting to the home page of Digg caused 20 bloggers to read about the site and blog about it, and that a mention by Gizmodo influenced 10 blogs posts, whereas a knitting site may get zero blog mentions from making the home page of Digg, but 30 mentions after being discussed in a crafting forum.

You probably can't determine the source for every link, but you can approximate pretty closely by looking at timing. And a lot of bloggers mention where they first saw a link. Do the same exercise for other types of awareness efforts. Which reporters picked up a story you sent them a press release about? Can you find patterns in the types of publications?

Similarly, make note of which links bring you the most qualified traffic. Which visitors buy something? Which referrals have the lowest bounce rate? Making the Digg home page may seem appealing if it brings you 5,000 vistors, but if the bounce rate is 95%, then it's less valuable than that crafting forum that brings you 500 visitors with a 15% bounce rate. (With those numbers, the Digg mention would be bringing in 250 qualified visitors, whereas the crafting forum would be bringing in 425.)

Knowing which efforts give you the most bang for the buck can help you know how to spend your resources next time.

TV-branding campaign

What should you measure?

It can be difficult to measure ROI on branding campaigns as they often aren't tied directly to sales. However, search data can provide clues about efficacy. You may be able to use overall search volume to determine if a campaign is increasing awareness.

A 2007 Jupiter Research study found that 2/3 of online searchers were driving to perform searches as a result of exposure to an offline channel. Raising awareness about your brand may cause more people to search for it. Use keyword research tools to monitor spikes in search volumes. For instance, notice the increase in search volume for [easy button] once Staples launched its campaign in January 2005 and the volume spike when it aired the related Superbowl ad in February.

Throughout the campaign, you can see several "breakout" searches (gaining more than 5000%) and related searches.

You should also make note of increased links and online chatter about your brand after you launch offline campaigns. This can be particularly useful if you have online elements connected to the campaign. For instance, if you are running a funny TV ad, you should make the video available online as well. You can measure chatter by doing things like setting up blog search and Twitter alerts, or by using a company such as Trackur or Visible Technologies to monitor things for you.

How is the data actionable?

You want to capitalize on your branding campaign as much as possible, so if it's influencing people to search for your brand or tagline, you want to make sure you can be found for those queries. Use the data from the keyword research tools to find out the top related searches and rising searches associated with the campaign.

In the case of Staples, their site ranks first for related queries such as [the easy button], [staples easy button], and [that was easy]. And keyword research (such as the Microsoft adCenter Add-In for Excel) shows that "the easy button" seems to resonate more with searchers than "that was easy":

Being ranked first is a great first step, but it's also important to consider the overall searcher experience. Consider this result for Nike's slogan, [just do it]. While the site appears first, the title and description don't provide a compelling marketing message to draw searchers in. If the site is ranking well but the traffic isn't what you expect, make sure that the display in the search results invites searchers in.

Another useful way to use search volume and inbound linking data is in comparing campaigns. Which caused search traffic to spike more and which brought in more links? Did having online elements to the campaign encourage more links and if so, which elements performed better than others? And as noted above, you always want to determine qualified traffic when looking at overall incoming traffic from search queries and links. Which campaigns caused search spikes that had the highest conversion rates?

Tracking online conversations can also alert you early in a campaign is creating negative perception, which can enable you to react right away and turn things around. Motrin, for instance, launched a "wearing your baby" ad intended to connect with mothers who had achy muscles, but instead, offended many of them, who shared their outrage with each other on social networks such as Twitter about their feeling that Motrin equated babies with fashion accessories.

Historically, these conversations may have taken place over the telephone or at dinner, and brands may have not known about negative perceptions like this until months later, after (expensive) follow up surveys and focus groups. But now brands that are monitoring the conversation online can learn about such sentiment in near real-time and can react accordingly.

There's lots of data out there. As a whole, it's likely too much to be useful. But by taking a close look at one or two key pieces of information, you can learn a lot about your customers and where to focus your efforts in the future.

Thanks Vanessa

Get Connected: Join us on our Facebook Page & follow us on Twitter: http://Twitter.com/adCenterBlog