Author Archives: Tara

Web Analytics: Capturing Valuable User Data

Posted on by
Reply

Anybody and their mom can copy and paste Google Analytics or Yahoo! Web Analytics tracking code onto their site. It takes someone special to be able to really drive insightful data with their implementation. Below are some steps you can take to help insure that your implementation is capturing the most valuable user data possible.

  1. Define your audience. If you have a travel site, who would want to buy your plane tickets? Maybe some of your travel packages are better suited to the elderly and retired, who can travel in the middle of the week. Identifying these audiences helps you think about how you might want to track them.
  2. Determine your signals of intent. What are the desired actions on your site? Form fills and phone calls have been the obvious conversion metrics for years, but don’t underestimate the value of “non-converting” web users. Say you are trying to drive a special vacation package to Florida, which comes at a discounted rate for AARP members. Some signals of intent might be viewing the deal, sharing it on Facebook (although this may be slightly less likely with AARP members), searching for flights and hotels within the package deal, using an online chat to ask questions, and so on.
  3. Track away! Now that you know who you’re looking for and what you want them to do, implement away! Here are some tips for the above scenario:
    • If the travel site has a registration or signup page which asks for gender, age range, ethnicity, religion, etc. you can set up custom events to “tag” your web users. You can then segment your data to look at your users that are 50 or older, Caucasian, Christian, or whatever your target audience is.
    • Yahoo! Web Analytics also offers valuable demographic data, which they capture via users who use Yahoo! products such as Yahoo! mail or Flickr.
    • Implement social analytics tracking for Google Analytics to measure who on your site is socially engaged with your content. You can monitor engagement of your travel deals pages over time to ensure your campaigns are encouraging organic promotion on social media networks.
    • Set up e-commerce tracking for your travel deal purchases, but also implement tracking to determine if users are even searching for a travel deals and packages – many web users will begin the signup process in order to compare prices, leave the site, then come back later to actually convert.
    • Some online chat services offer integration with web analytics tools so that you can get a little deeper info about chat usage (duration of chat) but if nothing else make sure to track clicks on the chat button or link.
    • When monitoring your web analytics, look at the full path to conversion and not just the last click. It may be that 70% of your conversions came from paid search, but there’s a good chance that display advertising influenced over 90% of them. Keep an eye on your conversion influencers is really key in determining how conversions are driven and, more importantly, how to allocate your budget and drive your web marketing strategy.

Any other favorite tips? Drop ‘em in the comments below!

May the Force be With Your Holiday Office Decorations

Posted on by
Reply

Our office is breaking out it’s holiday decorations, and with the typical nerddom that you would expect an internet marketing company to have, we are finding unique ways to incorporate elements one of our favorite science fiction series into the festoons.

These are some snowflakes that our handy Operations Manager created (can you name them all?):

During this cold winter season, one of the shadier characters at our office can be seen wearing this awesome X Wing Pilot Hoodie.

On our wishlists this year? These Yoda and R2-D2 Christmas lights…

Decking your halls with any nerdy boughs of holly? Please share them with us!

Quick Display Insights in Google Analytics

Posted on by
Reply

In honor of Display Month at Location3 Media, I’d like to share some tips on how to get quick insights on your display campaigns from Google Analytics.

In this specific example, I am looking at data for Display campaigns on the Google Display Network, but keep in mind that with some clever URL tagging you can do a similar type of analysis for Display campaigns purchased on non-Google networks as well.

To start, find your “Placements” report in Google Analytics. The below screenshot shows you where you can find this.

Once you are in the Placements report, you can choose to look at either your managed ad placements (which means that you are choosing which sites to display your ads on) or you can look at your automatic placements (meaning you allow Google to find relevant sites to display your ads on). In this example, we are looking at managed ad placements.

In the above report, you can see that placements like NYTimes.com (#1) earn the client a high amount of revenue, while placements like HomesForSaleinMA.com (#5) only earn a little revenue. But it can be hard to visually measure which of these placements are “better” based on this data. What we really need to know is, which placements have the highest revenue per visit?

To do this, we will quickly export this data to Excel. First, look at your “Show Rows” dropdown at the bottom of the page and extend it to show as many rows as possible.

Then, go to “Export” at the top of the page and choose the “CSV for Excel” version.

Once you have this open in Excel, you will need to do the following steps. This piece is fairly simple, but adding in screenshots for each step would be rather lengthy. If you need clarification on any of these, please let us know in the comments below.

  1. Use the “Text to Columns” feature of Excel to split the data into separate columns, using the comma as the separator.
  2. Add a column titled “Revenue/Visit”. For the data in this column, divide revenue by visit for each row.
  3. Highlight the table of data, then use the “Sort” feature of Excel to sort the entire thing by Revenue/Visit, with the highest value at the top (Largest to Smallest).

Now you will hopefully have a table like this:

Now you can see which sources are actually getting you more bang for your bucks! You can bid down on your more expensive, high-volume display sites and focus on these more niche sites that are performing well for your client.

Any thoughts/questions? We’d love to hear them!

Google Analytics: Flow Visualization Reports Revealed!

Posted on by
Reply

Today there was an exciting development in the web analytics world as Google announced a new reporting feature for Google Analytics: Flow Visualizations.

Two versions of this feature have been released, Visitors Flow and Goal Flow. The Visitors Flow shows the flow of your visitors from various web sources through your site, in a visually cognitive and intuitively interactive way. This is similar to Yahoo! Web Analytics’ Path analysis reports.

The Goal Flow functions in a similar manner to the old Goal Funnel Visualization, but with added fluidity in both appearance and functionality.

In both reports, you are able to achieve a much broader understanding of your web users’ navigation experience through your website than ever before. Bear in mind this feature will have a slow roll-out to all Google Analytics users, so you may have to have some patience before you can see this in your reports. These reports are only accessible from the new version of Google Analytics.

To learn more about this new feature, please read Phil Mui’s blog post on Flow Visualization on the official Google Analytics blog.

New! Google Analytics Multi-Channel Funnels

Posted on by
Reply

Attribution reporting is a crucial part of our web analysis process. Google Analytics announced today a new tool called Multi Channel Funnels, designed to help Google Analytics gurus attribute web campaign efforts, beyond just looking at last-click conversions.

Whatchoo Talkin’ ‘Bout?

Last-click conversions is a common term in web analytics that basically means that the last channel the user came through before making a conversion gets the “credit” for making the conversion. For example, if someone came to the website through Facebook, then through Yahoo! organic results, then through an AdWords ad and converted – the AdWords ad would be attributed to the conversion, even though Facebook and Yahoo! organic results clearly had an influence.

The new multi-channel analytics reports is a nice, visual way to attribute your conversions back to other campaigns. For example, it may be that your social media campaigns only had 3 direct or “last-click” conversions, but they may have influenced over 30 conversions. This is a much more accurate representation of brand impact.

What do you think? Any thoughts on how you would like to use this?

Forecasting With Excel Trend Lines: Pimp My Analytics

Posted on by
Reply

Every analyst has been there – you’ve put a lot of hard work into some insightful analysis for your clients, but they just don’t understand what you’re trying to say. How you visualize your insights can be equally as important as the wonderful data you’ve found. There are numerous wonderful data visualization tools out there, but the one I come back to time and again is good ol’ Microsoft Excel.

This is Part 1 of 3 in a series of some of my favorite tips for visualizing your data in Excel, specifically looking at forecasting, showing lost potential and measuring impact. Today’s post will cover forecasting.

For the purposes of these posts, I’m going to use a fake client, Tara’s Mortuary:

The Problem

One problem that I run into all the time is proving to the client the long-term value of a campaign. For example, say that I am using AdWords’ tablet targeting options to run a tablet-specific campaign for Tara’s Mortuary. I have seen in Google Analytics that tablet users convert well organically for my client, so I am fairly confident that they will convert well for my paid campaigns. However, my campaign is off to a slow start, with a very high Cost-Per-Lead (CPL) and a low Click-Through-Rate (CTR). I have been optimizing however (adding new ads to improve CTR, optimizing landing page to improve CPL), so these numbers are improving a bit. I need to prove to my client that with continued optimizations, we can get these numbers to their targets.

Trend Lines to the Rescue!

To do this, I will use trend lines in Microsoft Excel to forecast when we will hit the client’s targets.

This is how my Excel might be set up. Please note that I added more weeks than I have data for – this is in anticipation of my trend line. Without doing this, my forecasted weeks would not be labeled. Also, notice how I have added a value for the weeks I made optimizations – this is just to mark these dates in Excel. To mark dates in this manner, you will just need to use the highest max value for an axis on the date you want to mark. This may become clearer when you see the chart (below).

I added the drop-down lines for the optimization dates to my chart by using Error Bars. You can also see here the point of adding those Target CPL and Target CTR sets of data. I have also done a fair amount of formatting to this chart. The formatting is fairly intuitive so I won’t dwell on it here – but please ask if you have questions on how to do this.

There are a lot of different ways to add trend lines, but the easiest (in my opinion) is by selecting the line you want to add a trend line to, right-click, and select “Add Trendline…”

It will automatically open up the trend line formatting toolbox. There are quite a few options for trend lines within this toolbox.

For this example, we are going to use a simple linear trend line. However, it is a good idea to research and play around with the other trend line options. Here are some situations in which I use the other trend lines:

  • Polynomial: This shows the trend for lines that rise and fall. The “Order” is roughly the number of rises and falls in your data – for example, if your data rises, then falls, then rises again, you will want to set this to 3.
  • Moving Average: I use this often when I am looking at daily data, because daily data tends to fluctuate so much within a given week. Setting the “Period” to 7 will give an average over the course of 7 days, providing a nice trend line which has the same growth and decline as your daily data, without showing large rises and falls throughout the week.

As a rule of thumb I always give my trend line a custom name (selected in the screenshot above), and format the line color and style to make sense with the rest of my data. Below is the result!

I Got Your “Wow Factor” Right Here…

Thanks to our trend lines, we can prove to the client that if we continue the campaign through week 19, we will very likely hit our target CTR and CPL. When you do this, you may need to adjust the number of weeks you show based on when your trend line actually crosses the targets. You may not need the target lines or optimization dates at all – I find them helpful but sometimes they can be seen as clutter by the client, depending on how easily your client absorbs information.

What do you think? Have any favorite tips for showing future data to your client? We’d love to hear them!

Google Correlate: Insights on Your Audience

Posted on by
Reply

Sorry, Google Insights for Search – it’s not you, it’s me. I just think we should be spending time with other developers and web analytics tools. Lately, I’ve had my eye on Google Correlate. It’s new, which is exciting enough, but it is also incredibly powerful.

Correlate vs. Insights

Google Correlate works essentially the opposite way of Google Insights for Search. While Insights will compare queries that you select over time, Correlate will compare queries you select to other queries that have the same trend. For example, if we try to use Insights to compare “employment” and “real estate,” we will basically get a nice trend downwards for both, with news headlines along the way that might help explain rises and dips in the trend.

Very useful – but arguably not as powerful as Correlate. Insights (above) doesn’t really show us anything we don’t already know in this instance. Using Google Correlate, you can actually see which Google search queries have similar trends to “employment.”

This can help us come up with queries and ideas that we don’t already have – it’s a way to find the things that we “don’t know we don’t know.” However, this isn’t its most powerful feature…

Input Your Data to Learn About Your Audience!

Google Correlate has a feature that allows you to input your own data. If you don’t understand why this is exciting, here are some ideas to mull over:

  • Input visits from your clients’ social traffic going back the last few years and look to see what search queries have the same trend. You might see some great ideas for your social campaigns.
  • Input visits or conversions from clients going back the past few years by state (Yes, you can compare trends by U.S. state as well!) to identify regions that should be targeted by a local search campaign.
  • Upload all sorts of data from the U.S. Census, health records, driving records, you name it – there is a ton of great data sitting out there for free that you can use to make the most of this tool.

Our Example: U.S. Census Unemployment Rate by State

For one of our clients, we know that their target audience tends to be people who are struggling financially and may have even been recently unemployed. In order to learn more about this audience, we uploaded Unemployment Rates by State from May, 2011 as released by the U.S. Department of Labor, Bureau of Labor Statistics (whew!). The results that Google Correlate yielded were eye-opening.

Based on the results above, we learned the following about our audience:

  • Our audience is not only looking for basic retail jobs like Home Depot and Bath & Body Works, but they may also hoping for online jobs working for companies like MySpace.
  • Our audience probably spends a lot of time online educating themselves about different topics. They frequent sites like yahooanswers.com and wikipedia.org. This education may be related to finding a job.
  • Not only is our audience concerned with finding a job, they also have to grapple with things like writing hardship letters (something you have to write when you foreclose on a home) and getting out of debt. Knowing that these are potential concerns for our audience is huge – we can now target to help them address these issues as well as finding a job.

Your Turn.

This tool won’t achieve its full potential until we get out there and use it. How would you use this tool? Any ideas on how Google should improve it? Make sure to tell them here. I’d love to hear your thoughts on how we can use this tool – if you have an idea but aren’t sure how to get the data for it, we might be able to help.

Analytics, PPC & SEO: Finding Valuable Keywords

Posted on by
2

Using web analytics to supplement your PPC and SEO practices is fairly common practice nowadays. However, today I’d like to demonstrate an idea that came across my desk that is fairly new (at least to me), thanks to the innovations of our team here at Location3 Media.

Our goal is to find keywords via Google Analytics and other analysis tools with high value, low competition and a high traffic volume.

Calculating Value

To calculate the value of a keyword, we need to know how well it generates click-thrus and conversions. To the math-mobile! For our examples, we will use the Festa della Repubblica, which is being celebrated in Italy today.

Keywords Search Volume Visits Conversions ClickThru % Conv. % Value Index
festival of the republic 6,600 4,000 500 60.61% 12.50% 758
festa della repubblica 49,500 6,000 1,000 12.12% 16.67% 202
italian national holidays 10,000 5,000 900 50.00% 18.00% 900

To get the data:

  • Search Volume: Get this data from a tool like the Keyword Traffic Estimator for AdWords or Google Insights for Search. You are trying to get an equivalent of Impressions for this number.
  • Visits and Conversions: Just grab this from Google Analytics!
  • ClickThru %: Visits divided by Search Volume.
  • Conv. %: Conversions divided by Visits.
  • Value Index: This one is slightly tricky. Decide on a number of impressions that you would like to see for all keywords – it can be arbitrary at first, the number just has to be the same for all keywords. I used 10,000 for the example above. Then your equation is: [Impressions] * [ClickThru %] * [Conv. %] = Value Index.
    • So for “festival of the republic”…
      10,000 * 0.6061 * 0.1250 = 758

The beauty of this is that you would think that “festa della republica” would have the highest value because of its high search volume and conversions – but no! Putting more effort into keywords with a higher clickthru rate and conversion rate will ultimately yield you the best results for your investment.

Calculating Competition

For this part, all you really need is a great tool and the wherewithal to use it. I’m specifically talking about the AdWords Keyword Traffic Estimator tool in this example, as one of my favorite tools.

Looking at this tool in the interface will give you a handy visual of how competitive your keywords are:

However, we are number junkies, and this visual isn’t enough to feed our lust! Within AdWords you can export this to CSV for Excel and you will get a nice clean number:

The Result

So now we have Value and Competition, and we can pull in Traffic Volume from our earlier data:

Keyword Value Index Competition Traffic Volume
festival of the republic 758 0.04 6,600
festa della repubblica 202 0.01 49,500
italian national holidays 900 0.01 10,000

Now we have a great snapshot of our keywords! We can see that “italian national holidays” is our winner, because it has high value, low competition, and a high traffic volume. I would then place “festival of the republic” next on the list, because it has a high value, but I would keep in mind that the competition is higher for this keyword.

Some closing notes:

  • Make sure that all of your data has the same date range. Since AdWords gives you monthly estimated traffic, make sure that you are using monthly data from Google Analytics as well.
  • This is just one way to look at your data. Make sure that you consider all angles. In this example, it might make sense to still target “festa della repubblica” from a branding perspective, since that is the official name of the holiday.
  • You can pull all of this data from the AdWords interface as well, you would simply use Clicks instead of Visits, and Impressions instead of Search Volume. Keep in mind that your AdWords clicks will never exactly match your Google Analytics visits. In this case I used Google Analytics because it was relevant to an internal conversation here at Location3.

Thoughts? Ideas? What would you do differently? I’d love to hear it! Please leave your comments below.