Min-Maxing

Optimizing Emails for Blocked Images

2012-08-24 00:00:00 -0700

Emails that embed their content into images are not effective when a user's email client blocks images—and, while numbers vary based on source, roughly 50% of users (or their email clients) block images by default. By moving your content into text and styling your email to be effective in the absence of images, you can increase the likelihood that a user will take action based on your marketing efforts.

Why are images bad for email?

The lure of relying on images for content is a strong one. After all, the result is often attractive and successful at communicating that you represent a high-quality brand.

I recently had a chance to help a client (U.S. Patriot Tactical) improve their email template. This was the result:

The problem, though is that roughly half of email consumers use client that blocks image by default. As a result, this is what they would see in a worst-case scenario:

At this point, users have already made one decision to open your email, and with an image-heavy email you have now asked them to make a second choice: unblock images to see the content or ignore the message.

By placing another step between the user and your email content and giving them another opportunity to opt-out of consuming the content, you lessen the effectiveness of your email-based communication.

How can I improve the appearance of my email when images are blocked?

For the U.S. Patriot Template, we took some additional steps to improve their template code. This was the result:

To achieve this effect, we had to spend some extra time to perform the following tweaks to image areas:

Use alt text on images

The alt attribute on the img HTML element is used to specify default text that should appear if an image does not load for some reason (such as having the wrong image path or images being blocked by an email client).

<img src="http://path/to/image" alt="Text that will appear when the image is blocked">

In our example, we duplicated the content of each of the images as alt text for that image. For the social icons, we got as close as we could with alpha-numeric characters.

Use font and background styles on image elements

To style the alt text and color the areas where the images would normally appear, we used CSS styling on the img element. The following styles were used in this example in different ways:

    background-color: #72884E; // shades the image area
    color: #FC6; // controls font color
    font-family: Arial,sans-serif; // specifes the font style to use
    font-size: 16px; // controls the size of the alt text
    font-style: italic; // italicizes text to match the style of the logo
    font-weight: bold; // makes text bold

Keep in mind the following constraints when applying these styles:

Use in-line styles

Normally, you'd want to specify CSS styling in a separate CSS document (or, failing that, a style element at the beginning of your HTML document. However, email clients are notorious for ignoring some of all of your styles defined with these methods and overriding them with their own styles.

The only way to ensure that your image styles carry through is to specify styles using the style attribute on the image element:

<img border="0" alt="U.S. Patriot Tactical" style="display:block;font-size:20px;font-family: Arial,sans-serif;color:#222959;font-style:italic;font-weight:bold;" width="206" height="50">

When specifying fonts, use web-safe font stacks

Not all fonts are installed on all computers. If you have a specialized font that you embed on your site, chances are it is not installed on the user's computer. Since you can't embed fonts into an email, you should make sure that your alt text uses fonts that you know will be on most people's computers.

CSS Font Stack is a great resource for seeing the different options, learning how popular they are, and even copying code snippets for use in your emails.

Conclusion

Users spend on average just a few second scanning an email for anything that may interest them before deleting it. Anything you can do to make your content easy to consume in those few seconds will ensure as many of your users see your key points as possible.

How to Use SEO on Your Website

2012-08-11 00:00:00 -0700

You don't have to pay a firm thousands of dollars a month to apply the principles of Search Engine Optimization (SEO) to your site. All it takes is a little extra effort on the part of content creators and page template designers to increase your rank in search engine results.

Overview

The primary metric search engines use is the number of other high-quality sites that link to your content. The overall process for each piece of content looks something like this:

Post interesting content
Build your page in such a way that search engines can 'read' your content
Make your results appear attractive in search engine results
Use legitimate marketing means to spread word about the content

After that, it's up to other sites to link to you to help raise your relevance (and thus your rank).

How to Build Your Content

Identify content keywords
Write meaningful page and section titles using those keywords
Add the title of the page to the URL
Use meta tags and rich content snippets to improve appearance in search engine results
Make sure your pages load fast & are error-free

Identify Content Keywords

Since users discovering your site through search are doing so by searching for words or phrases related to your content, the first step to optimizing your site for search is to think about what words or phrases users are likely to enter into a search engine to get to content you write. For a product page, this may take the form of product names, brand names, nicknames, or key features. For articles, the topic, sections, and key concepts make good keywords.

Write Meaningful Page & Section Titles

Search engines give more weight to the content inside your page's header tags (<h1>, <h2>, <h3>) as these tags are meant to indicate the overall topic of the page as well as define the titles for sections. Using clear and meaningful page and section headers helps search engines to understand the overall concepts your page is covering.

In addition to the fact that search engines place some importance on the content of the <title> tag, this tag also controls the title of the page in search engines:

Note: Your content management systems (CMS) may handle the title and top-level header pieces for you when you enter the title of your page into the CMS.

Add Page Title to URL

Search engines give some weight to the text in the URL of a page as well, so placing the title into the URL as well can help rankings. In addition, users can look at a raw link and get an idea of what is behind it, which aids in overall understanding of what content hides behind it.

Note: Your CMS may also handle putting the title in the URL for you when you enter the title of your page into the CMS.

Use Meta Tags and Rich Content Snippets

By using the meta description tag(<meta name="description" content="Page description here">), you can tell search engines what you want the description of the page to be:

Keep in mind, however, that this doesn't guarantee that a search engine will use your description. If, for example, you searched for terms that appeared to be more relevant in the content of the page, search engines would use a snippet from the content instead.

Depending on the search engine, you can use additional markup to highlight additional information such as the author of an article and the date it was published:

Google has an article on how to show author information in Google search results.

Make sure your pages load fast & are error-free

To a minor degree, the time it takes for your pages to load can affect your page rank. While it only directly affects very slow sites, there is evidence that indicates users will spend less time on your site if it is slow. In addition, if your page has any errors, it can inhibit a search engine's ability to parse all of the content.

What about images?

The only clue search engines have for what your images contain is the alt attribute on the <img> tag. Make sure you desscribe what is inside the image using the alt text to make sure the image content is properly understood by search engines.

What about social media?

As of now, there's no huge argument for Facebook and Twitter affecting your page rank outside of marketing your content in hopes that people will link it elsewhere. Some search engines have begun looking at social networks for hints about popular content, but they only use it in limited situations. Search Engine Land has some more details on what social signals are used by search engines.

Bad SEO Practices

There are some methods people and SEO companies have attempted to use to 'game' search engine algorithms. In general, it's always a bad idea to attempt such methods as they are actively combated by search engine providers and they will penalize you if they detect that you use them.

Loading content via JavaScript after page load
Serving content to search engines that is different from what users see
Paid links & link spam
Duplicating Content

Loading content via JavaScript after page load

Search engine bots are like other bots in that they can parse the content that your page generates, but nothing that you inject after page load via JavaScript. Any such content will not be archived by a search engine.

Serving Different Content

Some people have attempted to increase the size of their page rank by stuffing keywords or extra content on a page such that normal users will not see them but a search engine would see them. Search engines are becoming progressively more able to detect this kind of content stuffing and will lower your rank if they detect this behavior.

Paid links & link spam

Since rank is dependent on how many sites link to you, some people have attempted to increase their rank by either spamming their links on other sites or by paying other sites to link to them. Just like with content hiding, search engines can detect spam patterns and will lower your rank accordingly. In addition, if there are links to your site from known 'bad neighborhoods' on the internet (link farms, for example), you will be penalized as a result.

Duplicating Content

Simply copying content from another site will not help increase your rank. Search engines look for duplicate content and penalize copycat pages. Google is even taking valid copyright claims into account in their algorithms: any site that habitually violates copyright laws may be penalized as a result.

Conclusion

Overall, once you lock in a page template that is friendly to search engines, the only on-going tasks you have is to make sure each new piece of content has appropriate keywords used for titles, headers, and throughout the content to help aid discovery via search. From there, keep writing interesting content!

How Limited Run Discovered Worthless Clicks Using Analytics

2012-07-30 00:00:00 -0700

A company called Limited Run made a post on Facebook about how they discovered that roughly 80% of clicks they received on their Facebook ads turned out to have the signature of a bot. They noticed that their site analytics reported a much lower user count from their Facebook ads than Facebook was reporting. After digging into the weird traffic by adding some custom tracking, they discovered:

The visits from those clicks did not load client-side JavaScript (hence why the site analytics didn't track the visit)
The visits were showing up as coming from non-standard browsers

This is an excellent example of the power of tagging the links you use in your marketing campaigns. Since Limited Run tagged the links that their Facebook ads used, they knew exactly which visits came from their Facebook ad campaign. That led to seeing that the numbers were much lower than Facebook reported and ultimately that most of the clicks were worthless.

Remember to tag your links!

Tracking the Results of Your Email, Social, and Ad Marketing Efforts

2012-07-28 00:00:00 -0700

Most social media and ad networks provide you with data regarding how your users initially engage with your marketing efforts. If you use an external vendor to send your marketing email, they will even go as far as telling you how many people open your messages and click on something inside of them. This information, while useful, only scratches the surface of the full impact your marketing efforts are having.

How can I determine which marketing efforts are driving traffic to my site?

The key to tracking this traffic is to tag links to your site with additional information at the end of each link. The format of this information will vary depending on what program you use to review the data. This tutorial will cover the format you will need to use if Google Analytics is your tool of choice.

What information can I tag my links with?

Medium: The format your marketing effort is taking.
Source: The company or site that is displaying your marketing to the user.
Campaign: The name of your marketing initiative.
Content: Optional. This helps differentiate multiple links to the same content on your site in your marketing efforts. For example, if you link to your homepage from a header logo and a footer link, you can use the content parameter to differentiate traffic that came in from those links. Another good use-case is differentiating clicks to a product through a product image, the product name, and a 'buy now' button.
Term: Optional. For marketing efforts where you are paying for specific keywords or terms, this will designate the specific term used for the link.

If you're having trouble building the links at first, you can use the Google Analytics URL Builder to get you started.

What are some examples of how I should tag my marketing efforts?

In the below examples, the pieces that change depending on the marketing effort are highlighted.

Email

Make sure to use the content parameter to differentiate which links are embedded in buttons, images, or text links. Knowing what draws the attention of your users and what drives the most site engagement can not only help improve your email marketing efforts, but your other efforts as well.

http://widgets.com/?utm_source=my-widget-company&utm_medium=email&utm_content=product-image&utm_campaign=widget-sale

Social Media

I recommend using the content parameter to help you uniquely identify which post on the social network had your link. This can help you connect the number of impressions, likes, shares, and other stats from that site with the number of pageviews and conversions on your site.

http://widgets.com/?utm_source=facebook&utm_medium=social-network&utm_content=2012-07-28-promo-with-image&utm_campaign=widget-sale

Banner Ads

The content parameter will help differentiate which ad units are getting clicked. You can use it to differentiate between medium rectangle (medrect), leaderboard, skyscraper, takeover, and other ad units.

http://widgets.com/?utm_source=awesome-ad-network&utm_medium=banner&utm_content=leaderboard&utm_campaign=widget-sale

Paid Search

The content parameter is most useful for indicating whether the cost method of paid search is cost-per-click (CPC), cost-per-acquisition (CPA) or something else.

It's vital that the company you are buying paid search from be able to supply the specific term as part of the link to your site. Knowing which keywords resulted in clicks is great, but knowing how much you made from those sales will give you the most accurate picture of whether your paid search campaigns are effective.

http://widgets.com/?utm_source=google&utm_medium=cpc&utm_campaign=brand-terms&utm_term=awesome-widget-5000

RSS Feeds

http://widgets.com/?utm_source=my-widget-company&utm_medium=rss&utm_campaign=my-blog-post

...and more!

There are many other places that you might link from if you stop to think about it. If you'll be linking to your site from a mobile application or game client, make sure you tag your link with a medium parameter so you can see how those links perform.

What do I do after I've tagged all of my campaigns?

After you've collected some data, you can go to the "Campaigns" report found in the side menu under "Traffic Sources > Sources". This report will allow you to filter based on your various mediums and can quickly show you which of your campaigns are generating users who stick around a while, perform favorable actions, and how much money they spent. Pretty neat!

A Word on URL Formats

The values you use for parameters are case-sensitive and character-sensitive. For example, the following items will show up as different mediums in Google Analytics reports:

social-network
Social-Network
social_network
social network

It doesn't matter which format you use. However, make sure everyone who is involved in tagging links uses the same format or you'll end up with messy reports that take you longer to use.

In addition, make sure everyone uses the same names for your campaigns as well. Otherwise, you might end up with a report like this:

Easy Date Range Selection with Qlikview

2012-07-20 00:00:00 -0700

By default, Qlikview allows you to use a date picker or slider object to select dates you can use as ranges to apply to a database field. There are a few problems with this approach:

Slider objects don't allow you to easily select specific dates when you have lots of possible dates.
Calendar objects don't allow for easy range selection across multiple months.
Using list selection boxes can flood the screen with lists of months and years.
By default, these date ranges will not affect charts based on different date fields.

Fortunately, it's easy to set up Qlikview with a more standard "select two dates" widget that will affect all of your charts, regardless of what database date fields drive their display ranges.

The process consists of three steps:

Define variables to hold your start and end dates.
Create calendar widgets that set the values for these variables.
Make your charts filter their different database date fields based on the values of these variables.

How do I set up variables in Qlikview?

In the 'Settings' menu, there's an item labeled 'Variable Overview...' which will allow you to define variables.

After clicking on the menu item, use the 'Add' button to define your variables. For this example, I'll use vStartDate and vEndDate.

How do I make my date selectors use variables?

On the 'General' tab of the 'Slider/Calendar Object', you'll see that the 'Data' option is set to 'Field' by default, this causing the object to affect the range of values for the chosen database field. Change this option to 'Variable(s)', then enter the name of one of the two date fields.

To limit the ranges a user can select, you can define a minimum and maximum value.

The cleanest way to do this is base the minimum and maximum of your start date selector on the minimum and maximum values that your database table holds with the Min and Max functions:

=Max(database_field_date)

For the end date selector, base the minimum value on the current value for the start date variable:

=$(vStartDate)

How do I make my date selectors affect multiple charts?

For each chart, go to the 'Expressions' tab in the properties, then click on the 'Definition' field and enter an expression to filter based on the variables your date selectors are now affecting:

=Sum( if( database_field_date >= $(vStartDate) AND database_field_date <= $(vEndDate), ga_visitors) )

The above statement will only select records where the database_field_date is between the two dates picked with the calendar selectors. For only those records, it will sum up the values in the ga_visitors field.

Of course, you can use any other aggregate function you choose. You can, for example, count distinct instances of users based on date ranges as well:

=Count( DISTINCT if( database_field2_date >= $(vStartDate) AND database_field2_date <= $(vEndDate), visitor_id) )

So long as the result of this expression is a number, your charts will be able to use it to display the appropriate data.

Once you've updated your charts, you are done! Simply adjust the dates using your selectors and watch as the charts update in real-time.

Importing Google Analytics Data into Your Data Warehouse with Pentaho

2012-07-15 00:00:00 -0700

Many companies that sell applications and games for desktop and mobile users build a data warehouse so they can create dashboards that show information about user patterns in these applications. By including web funnel data from your website into the data warehouse, you can go a step futher and get complete picture of your user acquisition and retention.

Prerequisites

Data Source

Our data source is Google Analytics. To access the data, we'll need access to the profiles via a username and password.

For the purposes of this project, I recommend creating a new Google Analytics account for this project and giving it read-only access to the specific profiles you want to retrieve. Since you'll need to enter this account's password in another piece of software, it's generally a better practice never to use your own accounts.

ETL Software

To get the data into your data warehouse, you will need software that can extract the data, transform it into the necessary format, then load it into your data warehouse. Such software is called Extract, Transform, and Load (ETL) software. For this project, I'll be using Pentaho Data Integration.

The ETL software will need to be able to connect to your data warehouse. You will also need the ability to create new tables and insert/update data into these tables.

Retrieving & Storing Data with Pentaho

Pentaho comes with a free connector that you can use to query data from Google Analytics. It uses the same syntax as the Google Analytics Core Reporting API to define filters.

Before you query Google Analytics, you'll need to grab the following information with the ETL:

A date range
The filters for each filter step you want

On top of that, we will hard-code the user authentication information, the dimension (time), and metric (visitors).

Job: Master

The main thing you should ensure is that the Get Web Funnel Data job is set to execute for every input row on the "Advanced" properties tab. This will cause it to execute for every row that is returned in the previous transformation where we look up the web funnel steps.

Beyond that, this job houses little direct logic. The notification emails are helpful if you schedule your ETL to run daily so you know whether or not it succeeded.

Transformation: Lookup Web Funnel Steps

To make the ETL more dynamic, it is helpful to store information about each funnel step in a database table. This will allow you to query for a new step in the future simply by adding a new row to this table. The table schema should look something like this:

funnel_step_id
- auto-incrementing sequence (primary key)
- Calling it "funnel_step_id" instead of "id" will make things easier for software that automatically links related tables together
name
- String representation of the funnel step for reports.
- Example: Landing Page Visitors
ga_filter
- String representation of the filter that Google Analytics will use to execute the query for this funnel step.
- Example: ga:eventAction==Start Registration

The "Copy Rows to result" output ensures the data is carried over into the next step of the ETL.

Job: Get Web Funnel Data

Like the master job, this one has little direct logic. It represents the loop that should execute for each funnel step that is returned from your earlier lookup.

Transformation: Set Variables for GA Query

This transformation is responsible for setting up the variables that you will need to feed into the actual query to Google Analytics.

Get rows from result

For this input, you should specify each of the table columns from your lookup table that you'll eventually set as an environment variable:

Get Start and End Dates

Because Google Analytics sometimes doesn't update their data for 24 hours, you want to be sure you're fetching data for the past few days so you can update days that might have incomplete data. The easiest way to do that is to use the start of last week for your beginning date and yesterday for your end date. This will mean you're always retrieving between 7-14 days of data.

Scrub Date Fields

Pentaho's date format differs from the format you have to use to query Google Analytics. You can use JavaScript to transform the data into the proper format.

var dateFormat = "yyyy-MM-dd";

sYesterday = date2str(sYesterday, dateFormat);
sStartLastWeek = date2str(sStartLastWeek, dateFormat);

Afterward, make sure you set up the fields to update with the new values in the table below the JavaScript text area:

Set Global Variables

Now that you have all the variables you'll need for the query, make sure you set them with a scope of "Valid in the root job" so that other transformations and jobs have access to the them:

Transformation: Query GA and Store Results

At this point, the ETL has all the data it needs. All that is left is to execute the query and store the results.

Get Visitors

Enter your user credentials, select the appropriate Google Analytics profile, then enter all of the static and variable information as part of your query definition:

Google Analytics will return several fields that you do not need. To remove them, select them in the "Output Fields" section, right-click on them, and choose "Delete selected lines" from the menu.

Get funnel_step_id

At this stage, the date and visitor count is part of your result set. You'll need to add the funnel_step_id to the result set so we can link this result to the correct funnel step in your lookup table:

Insert / Update

Now that we have the date, funnel step, and visitor count in our result set, we can insert the data into our data warehouse (or update existing rows):

In the first table, make sure to specify both the date and the funnel_step_id as the keys to use to look up the values. If Pentaho finds an existing record for that date and funnel_step_id, it will simply update it with the new results rather than add a duplicate entry.

Next Steps

In one of my next articles, I will show how you can take the information you have just imported into your data warehouse and create funnel visualizations using QlikView. This will allow you to connect the dots between all of your funnels and allow you to visualize the end-to-end flow of your users.

Multi-Channel Funnels in Google Analytics

2012-07-08 00:00:00 -0700

When reviewing the success of your funnels, one way you can segment your users is by where they came from when they arrived on your site. This can give you a better idea of which traffic sources are proving valuable to your efforts.

However, not everyone will perform a key action on their first visit to your site, and it may take some users several visits to reach the end of your primary funnel, and each visit can come from a different source. To assess the value of each of these sources properly, you'll need to understand multi-channel funnels and how to analyze them.

What are multi-channel funnels?

In marketing parlance, a channel is one avenue via which your site is advertized. How you categorize these sources differs depending on what's important to you, but some general categories for your marketing efforts might include:

Display (banner) ads
Web video ads
Promotional emails
Referrals from another site
Paid and organic search
Social media

Anytime you rely on multiple strategies to help drive traffic through a funnel on your site, you're using a multi-channel funnel.

What can I learn from a multi-channel funnel report?

There are several aspects to a multi-channel funnel that you may want to pay attention to when figuring out the value of each channel to your site. Different companies will place importance on each of these aspects differently depending on their use-case.

Last Interactions

The channel that drove a user to your site right before they completed a conversion funnel is flagged as a 'last interaction' channel. If a marketing effort has a high percentage of last interactions, it's helping to close the deal with your users. While this is the initial place most companies look when assessing the value of their marketing efforts, it doesn't paint the full picture.

First Interactions

By discovering what is driving a user to visit your site for the first time, you can discover which marketing efforts are helping spread initial awareness or interest about your brand. This will not allow you to easily differentiate between users who are just hearing about your site and users who have heard about it before and who simply saw something new that was convincing enough to check out your site. However, the value in either case is that a channel with a high percent of first interactions is clearly successful at driving initial interest in your brand.

Assisted Interactions

While both the first and last interactions are key conversions points, any channel along the way that engages the user will get credit for an 'assisted conversion'.

Many marketing efforts that have a seemingly-low return on investment (ROI) can actually turn out to be fairly important. You may have an email campaign you use to nudge new users to offer them help as they learn more about your site or product. This email itself may not convert a lot of users and it won't be how they discover your site, but noting that this effort assists in a lot of conversions will help identify whether this nudge email is truly succeeding at nudging users along your conversion funnel.

Popular Conversion Paths

This report will show which paths of first, assisted, and last conversions are responsible for converting the most users. While this information can be useful, the utility is somewhat limited. It helps to visualize channel funnels from a high level, but most of the actions you take will be related to first, last, and assisted conversions.

Time Through Path

One way to assess the success of marketing efforts is how much time passed between a user's first interaction and last interaction to conversion. You can use it to assess whether your efforts are resulting in people making decisions quickly. If you see that users are taking their time, you can use this information to experiment with ways to speed up their decision-making process.

How can I customize this report with my marketing channels in mind?

The default channels make a few assumptions about how you tag your marketing campaigns. When your existing tags don't match those assumptions, it can create a rather messy report.

Fortunately, Google Analytics allows you to override their default grouping and fix any issues you're encountering. You can do this through the admin menus, but I recommend starting by copying their default and editing it to meet your needs.

To copy the default grouping, click on "Channel Groupings" in the menu above the "Assisted Conversions" report and select "Copy Basic Channel Grouping template...".

On the next screen, you can give your new template a name. then, hover over any of the "Label Rules" to get an option to edit. This will allow you to see what rules are used to create these groups.

As you can see, my original example used an underscore ("social_media") whereas Google Analytics uses dashes or spaces in its original filter. I can easily edit the filter to include my variant so all of my existing campaigns are caught by the "Social Network" channel.

This screen also allows you to add new filters. For example, if you had a mobile app that has several links to your site, you might tag your links with a utm_medium parameter set to "mobile". Then, you can create a custom label rule to group all links with a utm_medium of "mobile" as a channel called "Mobile App".

Armed with a report customized for your website, it's much easier to make informed decisions on which of your marketing efforts are working and which ones need improvement.

Using Analytics to Find Actionable Data

2012-07-04 00:00:00 -0700

Pageviews don't make a site successful, but it's usually the first thing you see when you look at an analytics dashboard.

These stats are displayed so prominently that it's easy to think that they're key performance indicators (KPIs), and many companies treat them as such. The problem is that these aren't true KPIs; They're vanity metrics.

What are vanity metrics?

Vanity metrics are metrics whose changes don't directly correlate to increased conversions. While they might look pretty, they're only important in that they are the core building blocks for the data you really want to be analyzing. If you focus on vanity metrics, you cannot answer the questions you need to answer to make sure your site is healthy.

Why are vanity metrics dangerous to use as performance indicators?

Suppose I told you that my eCommerce storefront had 5% more unique visitors last month. At a very high level, that sounds great. Now, suppose I told you that revenue dropped 2% last month. Suddenly, it doesn't sound so good. Why am I making less money with more users?

Even focusing on overall revenue is dangerous. Suppose I said my storefront was making 5% more money. Cause for celebration? Not if I had 50% more new users during that period. Why did so many more new users result in such a low increase in revenue?

Each of these scenarios create more questions than answers. However, it's easy to dig past these vanity metrics and answer those questions with actionable metrics.

What are actionable metrics?

Actionable metrics answer specific questions about the performance of a site:

Am I making money on my marketing efforts?
Where are visitors dropping off before they convert?
Is there something I can do to make fewer users drop off?
What is likely to cause users to convert at a higher rate?

Note that these questions all focus not just on raw numbers, but attempt to get answers that can result in direct action to improve the site.

In my previous example, I mentioned a site that saw 50% more new visitors but only 5% increased revenue. To figure out why, a good step would be to see how those users progressed through the purchase funnel.

Now, the problem in the example is a bit more clear. The ad landing page had a lot of traffic that never got to the product page. The product page was pretty successful in that most people added the product to the cart, but most of those people didn't follow through with the purchase. Armed with this kind of information, it's easy to identify part of the site that should be considered for split-testing. With some tweaks, it's possible to reduce the friction of this funnel and end up with more sales.

You'll have to dig deeper for actionable metrics than you do for vanity metrics, but it's worth it. Fortunately, most modern web analytics solutions provide standard reports that help discover many common actionable metrics pretty quickly. For information that's specific to your organization, you can also create custom reports that you can refer back to for tracking the progress of important actionable metrics.

Turning Weight Loss into a Game

2012-06-24 00:00:00 -0700

Losing weight is not fun. At some level, you're forced to remove parts of your diet and lifestyle that are familiar and comfortable. In their place, you adopt things that are boring and difficult. People who are successful at losing weight tend to find different ways to make the process interesting by finding a sport they like to play as part of their exercise or healthy foods they like to eat.

For a while, I played basketball with friends and switched to some diet versions of foods I normally ate. This was enough to help me drop from my all-time high of 235 lbs. to 195 lbs. Then, my weight loss flat-lined. The problem was that I was still intaking enough calories that, even with the burn of exercise, I'd reached equilibrium with my metabolism. I needed to cut calories to get any further, and that seemed like a great way to annoy me and ultimately derail from my plan.

I needed a way to get over this hump while making this next phase of my weight loss interesting enough to pursue.

Research

I started my venture by finding answers to the following questions:

How many calories equal 1 pound of fat?

1 pound of fat represents 3,500 calories^[1]. To lose a pound of fat, your body has to have a net loss of 3,500 calories over the course of a period of time. Spoken differently, you have to use 3,500 more calories than you ingested.

How many calories do I burn a day?

This question consists of two parts.

How many calories do I burn at rest?

Your Basal Metabolic Rate (BMR) is the rate at which you burn calories at rest. This differs from person-to-person, but fortunately you can use a BMR calculator to figure out how many calories you burn in a day.

At 198 pounds, I burned 1922 calories while at rest.

How many calories do I burn doing my normal daily activities?

The Harris Benedict Principle is a rough guideline for figuring out how many calories you burn performing your daily activities. While not an exact number, it will give you enough actionable information:

Activity Level	Example	Calories Burned
Sedentary	desk job, no exercise	BMR x 1.2
Light	regular walking, exercise 1-3 times a week	BMR x 1.375
Moderate	somewhat physically-demanding job, exercise 3-5 times a week	BMR x 1.55
Heavy	physically-demanding job or daily exercise	BMR x 1.725
Very Heavy	physically-demanding job & daily exercise	BMR x 1.9

I burned 2306 calories in a day at 198 pounds while working a desk job.

At what point does weight loss become unhealthy?

If you intake a low amount of calories, you can trigger your body's starvation response. This puts your body in a state where it breaks down non-fatty tissue. This not only makes it harder to lose weight, but it decreases your overall health.

To avoid the starvation response, you should end up with a net caloric intake of around 20% lower than the number of the calories you normally burn in a day. For those with a higher percentage of body fat, this can go up to 30%^[2].

I ended up creating a profile at My Fitness Pal to help calculate a healthy weight loss rate. As part of the process, the site had me enter my current and goal weights, and then gave me a per-day calorie goal. That number started out at 1640 calories per day, which is about 29% lower than may daily calorie burn. This equates to losing about 1.3 pounds a week.

How much should I weigh to be healthy?

Your Body Mass Index is a measure of the amount of fat in your body. You can use a BMI calculator to determine your BMI and see the normal range for a person of your height.

For a 6'1" male, a healthy weight is between 140 and 190 pounds. I decided to split the difference and make 165 pounds my goal.

Execution

Having a set of rules and numbers to work against suddenly made the concept a little more interesting to me. It was beginning to turn weight loss into a game. By using My Fitness Pal (MFP) to track my caloric intake each day, I could see my daily progress toward my goal.

MFP has a large database of foods already entered by site administrators and users, so tracking most of what I ate at home and restaurants turned out to be pretty easy. Anything that was not in the database was easy to add through their interface. I ran into a few circumstances where I had to enter nutritional information into the site by copying values over from restaurant nutrition guides or healthy recipe sites, but even that was pretty easy. MFP also allows users to group foods together as "meals", so I could easily take the bread, mayo, cheese, and meat for my sandwiches and combine it into one "meal". Then, I merely had to add a serving of that meal when I wanted to use it.

When I fell into the practice of eating the same healthy breakfast over and over, it became even easier. All I had to do was copy my "breakfast" menu from the previous day.

In one case, my wife had a good recipe for a healthy mexican lasagna, but no nutrition information for it. After entering the ingredients into a spreadsheet, however, I was able to calculate the nutrition content per serving. I entered that as a meal. Shortly thereafter, MFP released a custom recipe feature that facilitated doing this without having to use separate spreadsheets.

As I started tracking my caloric intake, I became fixated on other numbers. The Andriod application for MFP allowed me to see how much of the various nutrients (fiber, potassium, sodium, various vitamins) I was intaking on a normal basis as well. I started structuring my diet to intake the proper amount of calories while ensuring I reached my good nutrient goals as well (ideally via food, but also via dietary supplements if I was under a vitamin goal for the day). Tracking the various numbers during this process was the final step in turning my weight loss goal into a game for me, and before I knew it I was well on my way to my goal.

Tracking my protein and fiber intake proved to have an additional benefit: protien and fiber are great ways to make food "stick with you" (basically, keeping you feeling full for longer without artificial appetite suppressants).

Takeaways

Turning weight loss into a game was ultimately the change I needed in my diet plan to get me to reach my goal weight. The game became the core focus, and everything else I did (which normally would have felt like sacrifice) was a way of min-maxing my nutrient stats to win the game.

I still ate out but I was able to find healthy alternatives that were pleasing without feeling like I was forcing myself to eat something I didn't really like for the sake of health. The only time it got annoying was when I had to enter my chosen meal as a custom food in MFP because it wasn't already there, but that only happened occasionally, and even then it was only early-on, before I had entered all the foods I eat regularly when I eat out.

I did exercise a bit throughout this process, but it wasn't the main factor in my weight loss. In fact, if I worked out during one day, it just meant that I had to eat more that day to make sure I kept above the level via which the starvation response was triggered. In the end, exercise was a great way to treat myself to a dessert or a less-healthy meal. Of course, working out is a very good addition to a weight loss plan for the other health benefits as well.

Obviously, different methods will work for different people with regard to losing weight. Physicians and personal trainers can offer better advice than most people (including myself) with regard to tailoring a healthy weight loss plan to your needs. If you can, though, find a way to merge it with one of your passions.

Min-Maxing

Optimizing Emails for Blocked Images

Why are images bad for email?

How can I improve the appearance of my email when images are blocked?

Use alt text on images

Use font and background styles on image elements

Use in-line styles

When specifying fonts, use web-safe font stacks

Conclusion

How to Use SEO on Your Website

Overview

How to Build Your Content

Identify Content Keywords

Write Meaningful Page & Section Titles

Add Page Title to URL

Use Meta Tags and Rich Content Snippets

Make sure your pages load fast & are error-free

What about images?

What about social media?

Bad SEO Practices

Loading content via JavaScript after page load

Serving Different Content

Paid links & link spam

Duplicating Content

Conclusion

How Limited Run Discovered Worthless Clicks Using Analytics

Tracking the Results of Your Email, Social, and Ad Marketing Efforts

How can I determine which marketing efforts are driving traffic to my site?

What information can I tag my links with?

What are some examples of how I should tag my marketing efforts?

Email

Social Media

Banner Ads

Paid Search

RSS Feeds

...and more!

What do I do after I've tagged all of my campaigns?

A Word on URL Formats

Easy Date Range Selection with Qlikview

How do I set up variables in Qlikview?

How do I make my date selectors use variables?

How do I make my date selectors affect multiple charts?

Importing Google Analytics Data into Your Data Warehouse with Pentaho

Prerequisites

Data Source

ETL Software

Retrieving & Storing Data with Pentaho

Job: Master

Transformation: Lookup Web Funnel Steps

Job: Get Web Funnel Data

Transformation: Set Variables for GA Query

Get rows from result

Get Start and End Dates

Scrub Date Fields

Set Global Variables

Transformation: Query GA and Store Results

Get Visitors

Get funnel_step_id

Insert / Update

Next Steps

Multi-Channel Funnels in Google Analytics

What are multi-channel funnels?

What can I learn from a multi-channel funnel report?

Last Interactions

First Interactions

Assisted Interactions

Popular Conversion Paths

Time Through Path

How can I customize this report with my marketing channels in mind?

Funnels

Segments

Using Analytics to Find Actionable Data

What are vanity metrics?

Why are vanity metrics dangerous to use as performance indicators?

What are actionable metrics?

Key Performance Indicators

Conversion

Funnel

Split-testing

Turning Weight Loss into a Game