Geary Fresh

Behavioral Targeting: A Web Analytics Approach to Segmentation- Part 2

In my prior post, I discussed using segmentation as a way to better understand key metrics by removing the noise of your non-targeted site audience. For this post I would like to discuss some guidelines for identifying segments and targeting them to help improve key performance indicators.

By using Web Analytics data, you can identify different segments based on specific actions that your site visitors are taking once on your website. Trying to remain broad here as there are many different types of websites; any action a user takes with your website, through a call center, or in-store, defines an opportunity to identify a behavior. That behavior can provide some very valuable information that can benefit marketers with re-targeting strategies.

For example, let’s say that you have a content site with lots of information that is separated into different categories. You could target your returning visitors with ads that are relevant to their interests based on the different sections of the site that they visited.

 Another example would include someone who visits a commerce site repeatedly because they are in the research phase of learning about one of the products that you carry. After the second or third visit to a specific product page, you could serve them a type of incentive to persuade them to purchase at that time. I hope your competitors haven’t thought of that yet.

Here are some guidelines to help with your behavioral targeting strategy:

1.       Clearly define the goals of your website

2.       Identify key segments (Note that additional research aside from web analytics data may be necessary to define)

3.       Understand the goals for each of your segments and the increased value they will provide

4.       Select the appropriate tools that will allow you to target your site visitors

5.       After implementation, analyze your segments and identify improvement opportunities

Behavioral targeting is not a one time initiative; a continuous improvement plan should be performed as segmented behavior may change over time depending on your website goals.

No, Social media isn’t TV. And you think that’s a bad thing why?

‘Marketers’ still have some thinking to do about our future.  And for all that we tend to prattle on about the customers coming first and respecting our ‘targets’ as people first, purchase-prone message-victims second, I suspect that deep in the collective marketing psyche we’d be happier if they’d just follow the rules we learned in grad school and respond appropriately to reach and frequency. Buy what we tell you, because we told you to.  And we have research to back that up.  So what’s your problem, Mr. Customer?  (Who am I kidding; in CPG land that would be ‘Ms./Mrs. Customer’.  Viva la (gendered) purchaser.)

As you may gather from my  tone, I am not a fan of this approach. And I’m struck by the paradox inherent in MediaPost’s article relating the findings of the new Knowledge Networks report (Check out the link to the mediapost article here:  http://www.mediapost.com/publications/?fa=Articles.showArticle&art_aid=106445) that ’social media is ineffective re: advertising’.  For ‘advertising’ as we old reach-n-frequency folks have known it, I’m sure that’s true.   But the Knowledge Report study finds that good ole’ Word of Mouth is still the #1 promotional force in consumerland.  What’s word of mouth?  People talking.  TRUSTED people talking to each other.  And what’s social media?  The same bloody thing.  Advertisers, to be sure, will not be the main thrust of those many, many conversations, and some who try too hard to insert themselves into organic conversations come across as either boorish or clueless.  That’s the reach and frequency game.  Conversations – real ones – are different.  But I challenge the notion that social media isn’t an effective playground for smart marketers.   Word of mouth has always been trickier – and more demanding of actual human understanding, goals and motivations – than the broadcast model.  Why was anyone expecting social media to be different?  Don’t write off the medium because you expected it to function like a TV set.  Rather,  start looking at your customers are people.  They’re not always buying your product, but they are always being themselves.  How can you be relevant to them?   Actually relevant, rather than ‘relevancy powered by quantitative research points’?

Isn’t that the beginning of any good conversation?

- Sarah

What do customers want? A look at desirable brand traits.

eMarketer reported today a list of the qualities customers respect in a brand. Conducted by MTV, the survey looks at the 18-24 age group to determine if a brand is indeed king. It turns out that a label is not the almighty persuader we might think. The top 3 most important characteristics are “good quality, trustworthy, and works well.”

Here are the results,

• 74% good quality
• 55% trustworthiness
• 51% the product works well
• 46% it has a cool image
• 46% it has a good history
• 46% popularity
• 44% knowing it has been good for a long time
• 40% it has class
• 39% fits with my own image
• 33% its authentic
• 33% uniqueness
• 32% people are talking about it
• 31% its cutting edge
• 29% it fits in with other products
• 25% friends like it

What’s most important to you when making a purchase?

Remove the Noise: A Web Analytics Approach to Segmentation- Part 1

Segmentation can be used to analyze site visitors in a multiple of ways to compare different types of behavior and identify new insights. Another method of segmentation that can be used is to eliminate visitors by removing the noise of the un-targeted traffic that is not intended for your website.

For example, hopefully you currently exclude Internal IP addresses so that your employees’ activity on the website is excluded from reports. However, if your site only sells products or services in the U.S. and you do not provide services in another country; then you might say that your analysis may be a better representation of your site visitors or intended audience if you filter our traffic from other countries.

Some may argue that having that data is valuable as it may provide insight to the demand of your product or service in another region. This is true and thus should be the reason why you always have a profile that collects all web data to your site. It’s a lot easier to segment out the data you need to do an analysis than it is to add it back in (if that is even an option).

So what are some other ways you would find value in segmenting out types of behavior to better understand your target audience?

Twitter: 10 things to remember before a tweet

1. You have an audience. Even if you’re not following someone’s updates, they can still see what you post. If it’s not something you would say to your boss, don’t say it.

2. You cannot undo a post. On top of that, Twitter posts are crawled by Google. By no means do these facts call for censorship, but a second of pause before submitting a post does not hurt.

3. People follow updates to hear information that appeals to them. If you are managing a corporate/ branded account, stick to appropriate topics and don’t veer too far off course. If you do want to broach a new topic, take it slow and give your followers as much context as 140 characters allows.

4.  Frequency matters. Ever notice that you start glossing over updates from a user if they over post? While some leaders of the Twitter pack can get away with 30+ posts a day, stick to updates that people will want to pay attention to. That way…they will.

5. Tone matters. While 140 characters is not a ton of space, it is sufficient to establish a tone and POV.

6. Diction is your friend. Saying what you mean to say can be difficult. On Twitter every word counts. Make them count.

7. People like to be re-tweeted. It makes them feel worthy. If you don’t have something to say, find someone who does and give them credit where credit is due.

8. Monitor when you gain followers. If you send out a post about what you did last night or what article you’re reading…keep it up. While its not an exact science, this measurement will show you what fellow Tweeps are interested in.

9. Follow back. No one likes to dance by themselves, so if someone follows you, check to see if you’d like to return the favor. Do not feel obligated to follow everyone, but it doesn’t hurt to look. If there interested in what you have to say, the opposite is likely true.

10. Lastly, remember Twitter is social, interactive and communal. If your goal is to connect with your consumers, provide them value. It’s possible to communicate without constant promotion–in theory at the very least.

Social media marketing: What’s legal and what’s not?

With all the hubbub surrounding social media marketing, it’s easy to check your ethics at the door. Remember when Wal-Mart got caught astroturfing – that is, setting up a a corporate blog that masqueraded as grassroots buzz? The blog, Working Families for Wal-Mart, earned Edelman – the public relations firm responsible for it – a suspension from the Word of Mouth Marketing Association.

Underhanded tactics like astroturfing aren’t just unethical, however. In some cases, they are illegal.
There are numerous FTC regulations surrounding social media practices, including buying links, giving bloggers freebies in exchange for writeups, and viral content. In addition, Google has its own standards on what’s ok and what’s not.

Of course, these regulations are notoriously difficult to enforce. But, before you begin any social media initiative on behalf of your product or company, it’s worth it to understand what’s truly illegal, what’s simply underhanded, and how you can embrace social media without abandoning your ethics.

Read the full article, “Is Social Media Marketing Illegal?“, at SEOMoz.

Google Chrome Viral Videos

The launch of Google chrome was relatively low-key. It went live, media outlets mentioned it–that was about it. Now, Google is stepping it up a notch. Today the search engine Goliath launches 11 YouTube videos in the hopes that they will become “viral.”

This blog post proves that their move is working.

The videos do not call out Chrome’s competition, namely Firefox, Internet Explorer and Safari. Instead, the campaigns encourages users to think about what they really want from a browser. Each video varies drastically from one to the other, but all in all, they tout that Google Chrome improves online efficiency and general snazziness.

To see all of the videos, visit http://www.youtube.com/user/googlechrome. It’s definitely worth the trip.

Coca-Cola Promotes Performance Driven Marketing

Coca-Cola Co. announced its plan to focus on performance at the Association of National Advertisers Financial Management Conference in Phoenix on April 20. As noted in an Age Age article published this morning, Coke’s performance-driven model has been successfully rolled out in 5 markets. Plans currently include to expand this pay-for-performance marketing model to 35 additional markets within the year.

Coke urges other agencies to adopt this foreward thinking engagement structure to ensure that clients pay for Real Results–as opposed to typical hourly rates. Ad Age states that “The concept of value-based models has been a hot topic in the industry for at least a decade, but few marketers have attempted to apply it. (Procter & Gamble is perhaps the furthest along but uses it on only 12 of its brands.)”

Numerous service-based companies are incorporating this structure into their contracts, but some naysayers believe changes made now are just reaction to the recession. Whatever the catalyst may be, the move to client-focused pricing will ensure agencies are accountable for the success of their projects. If results exceed expectations, agencies and clients will win.

Original Ad Age article:

http://adage.com/article?article_id=136266

Other columns to add to the discussion:

http://www.theaustralian.news.com.au/business/story/0,28124,25347479-36418,00.html

http://pm2pm.blogspot.com/2009/01/advertising-agency-compensation.html

Packet Sniffers – The X-Ray Glasses of Web Analytics

One tool in a web analyst’s belt that is critical in the implementation of a web analytics solution is a packet sniffer.  A packet sniffer allows you to view the data that is being collected from your site and sent to the analytics tool’s servers.  This tool allows you to not only troubleshoot your implementation by viewing that data is currently sent, but it gives you a ton of insight as to what behavior is dependent to the type of data that is collected.    

I understand that packet sniffers have been criticized for displaying too much information, or that you must be of the “technical type” in order to decipher the image request.  There are plenty of packet sniffers available that offer a diverse amount of features that anyone can use.  In order to ease the learning curve, and to convince those of you who might be reading this with a raised brow, here are some tips that I’ve learned during the years:

Get the right packet sniffer for you – I always hear the argument that the packet sniffer is too complicated.  There are packet sniffers that offer as little as the actual request from the analytics tool or as much as the “kitchen sink.”  The idea is to match your needs with the capabilities of the tool.

Find a list of variable parameters – Another point of confusion when reading data from an image request is that the parameters don’t match the variables in the code.  You can find list of parameters with the variable that it matches from the vendor.  You can use the list to familiarize yourself with the data that is collected when you travel from page to page on your site. 

Learn how to filter your requests – Each analytics tool uses a unique identifier in their requests that can be used to filter them from other requests.  Here’s a list of the most popular tools:

·         Omniture SiteCatalyst – b/ss

·         Omniture HBX – HG

·         Google Analytics – utm

·         Webtrends – /dcsd

·         Coremetrics – /cm?

A packet sniffer not only displays what information is sent from your site, but also displays other information that can help you learn more about how your analytics solution behaves and how to improve it.  The answers might not be as polished as you may like, but that’s part of the fun.  So, what do you think, have I convinced you or confused you?

Is the Associated Press a Google-killer?