1. Traditional web analytics

Thanks to Google Analytics’ free-ness, no one has an excuse to be without web analytics! (Going without a web analyst is another story). These tools are essential for understanding where visitors come from, what users do on your site, what content is most popular / effective.

I’m vendor agnostic when it comes to analytics tools, but I do recommend everyone have a Google Analytics account for a few reasons. Besides the fact it’s free, GA integrates with Google Adwords and Google Website Optimizer nicely. There’s also several books and blogs entirely dedicated to Google Analytics, so there’s a wealth of resources for using it.

However, Google Analytics may not be enough if you have advanced analytics needs, such as importing cost data, for example.

Even if you’re using a Gucci paid analytics tool, it’s a good idea to collect data with Google Analytics as a backup (if anything ever goes wrong with your enterprise tool, you won’t have periods with missing data.) You also won’t have to worry about losing historical data if you want to switch vendors down the line, as some vendors will not let you keep your data or port it over to your new tool.

2. Optimization testing

Again, no excuses! Google Website Optimizer is a free tool that can get you up and running with A/B or multivariate testing in no time. There are also paid site testing tools worth checking out, like Optimizely, Unbounce and Visual Website Optimizer. Check out WhichMVT.com for a comprehensive comparison of site testing tool options.

3. User testing

User testing has a reputation for being expensive and time consuming, but it doesn’t have to be. You can get good, qualitative feedback on user experience for $40 a test with Usertesting.com. Since you can get good data from as little as 5 participants, $200 can get you started.

Tip: use your user test data as a springboard for A/B and multivariate tests.

4. Customer surveys

Like user tests, customer surveys provide great qualitative data. But even better than user tests, they provide feedback on how your actual customers feel about your site.

There are several “voice of the customer” tools you can use to collect data, including OpinionLab, iPerceptions’ 4Q product, Kampyle and even Survey Monkey. Many of these tools will integrate with your web analytics package. Some, like OpinionLab, also have testing capabilities.

Check out related articles Tips for Customer Surveys and Creative Places to Ask for Feedback.

5. Webmaster tools

Both Google and Bing offer webmaster tools that will tell you general health of your website from search engine’s perspective – broken links, pages not being crawled properly, errors in sitemaps, etc. Hey, if search engines offer this information for free to help your SEO, you better take it.

6. Brand monitoring / reputation management

Reputation management means monitoring and responding to mentions of your personal name, brand name, product names and other trademarks online.

At minimum, you should have Google Alerts and Twitter alerts set up.

Google Alerts is a free service that lets you track mentions of your brand name, personal name, product names or competitor products and trademarks. You can receive alerts in real-time, daily or weekly.

Tweetbeep is like Google Alerts for Twitter, and it’s awesome (when it’s working). I use the free version, but because it works intermittently, I also subscribe to mentions through “saved searches” feature in Twitter. You can also subscribe by RSS to any Twitter search, so they appear in your feed reader.

Some Twitter clients like Hootsuite allow you to to subscribe to any search term:

For real reputation management analytics, you’ll want a tool that can extract insights from this data, such as trend graphs and measure of influence, like Trackur.

7. Social analytics

In addition to just listening to what people are saying about you through the Web and social media, monitoring the traction your social media attracts is important. This means tracking engagement in your on-site forums, Facebook Pages, Twitter streams (follows, unfollows, retweets) and so on.

Facebook has its own analytics tool, and many third parties can track engagement on Twitter in a “social media dashboard.”

The Radian6 platform can even import web analytics and CRM data so you can track back ROI to campaigns and social activities, or capture conversations and social media profiles and tie them to customer accounts.

8. Eye Tracking

For $49 per test, Gazehawk will show your URL to real users and track their eye movements with webcams (which is why the pricing is so reasonable compared to fancy equipment lab testing).

The big bang you get from eye tracking data is the ability to see what people look at (eye fixations), and compare it across test versions of your site. While you can run A/B tests and measure home page bounce rate and clicks on certain links and features, eye tracking heatmaps show you what parts of the design got the most attention, and how behavior may change according to design.

See how ads on Virgin Megastore changed the way visitors “saw” the entire site. The bold red sale banner actually resulted in more navigation menu attention, while a green banner attracted more fixations on itself.

Photo credit: Etre

Combine eye tracking tests with A/B or multivariate tests and you’ll be able to answer more questions on “why” you got certain results than simply taking quantitative data and trying to explain it. (E.g. cart abandonment high because customers aren’t “seeing” the guest checkout option).

9. Click tracking

Click tracking heat maps are another way to visualize the effectiveness of your design, but rather than eye fixations – you got it – you’re looking at mouse movements and clicks. (According to Click Tale, independent research shows that there is an 84% to 88% correlation between mouse and eye movements).

Though you likely have a site overlay feature in your web analytics (Google replaced Site Overlay with In Page Analytics a few months ago), tools like Click Tale and Crazy Egg have advantages.

For example, one of the biggest downsides of Google’s tool is if you have multiple links to a URL on a page (say, top navigation and a merchandising zone), it will combine and display clicks from both links – you’re not sure which link is used more. You can workaround this with URL parameters like “top_nav” or “hp_banner,” but keep in mind that can complicate your web analytics, as each URL will report its own stats.

Another benefit of click heat maps is you can see what a customer clicks on, even when it’s not clickable. You may discover un-linked images get a lot of clicks, which can frustrate users who think your site is not working. For example, a software company that posts screen shots that include call to action buttons. The image, of course, is not clickable – but calls to action in the screen shot may get more traction than the landing page’s!

Similar to click tracking is mouse tracking, which shows you where and how long users hover. This is helpful to determine whether customers use your tooltips, AJAX goodies, and other unclickable content. (A longer hover may indicate the customer is actually reading your content.)

10. Mobile

If you’ve invested in a mobile site or app, mobile analytics can’t be ignored. Your traditional desktop analytics tool most likely reports mobile stats, but that data can’t completely be trusted. There are several problems with traditional analytics for mobile:

• Javascript tagging is poorly handled by many devices, and is not very reliable
• Cookies are often deleted after mobile browser closed or when phone turned off
• Unique visitors ID come from the carrier’s gateway, which under-represents unique visitors, and mis-reports locations (all your Nokia visitors coming from overseas, for example)
• Not all devices (especially older ones) are supported by desktop analytics tools
• Some (like Google Analytics) use sampled data, which isn’t 100% accurate
• Mobile apps have unique metrics that desktop analytics won’t report

For businesses whose mobile apps are the product (think of mobile versions of software like Microsoft’s Documents to Go and Documents to Go Premium), mobile analytics that show conversion rates for free trial to paid upgrades, upgrade revenue, time from free trial download to upgrade and uninstalls is extremely important. Traditional analytics can’t handle these emerging metric requirements.

Mobile analytics vendors Bango and PercentMobile have specialized in solving these m-analytics issues, and can offer you a much more accurate picture of true mobile usage and success metrics.

11. Competitive Intelligence

CI is perhaps the toughest to attain, but there are tools out there that can attempt to answer the question of how you stack up to the competition in traffic, sales, conversions, growth, search ranking, search popularity and so on.

A few freebies exist from Google, of course. Google Analytics evangelist Avinash Kaushik has described the virtues of Google Insights for Search, Google Trends for Websites and Google Analytics Benchmarking on his blog.

Paid tools include SpyFu for PPC campaigns and keyword research, Compete and Hitewise for search and site behavior, and Comscore for online advertising.

Hitwise and Compete provide in-depth data which many can’t afford. Another tool to consider is Beencounter which tracks what percent of your visitors have also visited your competitors’ sites (or any other sites you’d like to track). This information has applications in advertising planning and web personalization.

More important than data…

This list of data types may overwhelm you, especially if you’re struggling just to get good at traditional web analytics as an organization. As many wise have said, tools should make up 10% of your analytics budget – people should make up 90%. Remember that to the degree you expand your analytics horizons is the degree you need more good analysts on your team.

On December 29, (the same date of the Headsets mail) I also sent out a non-marketing, “hey thanks” letter to our Gold Guys Canada base. I was motivated to do so because a) I wanted to remind our clients that we are individuals behind the face of our business, b) I wanted to reconnect on a personal level with our clients and c) I was tired of always crafting a “pitch” and relished the idea of a gentle, human letter.

I don’t know if the Headsets results were the same as ours, but we saw:

• A 7% drop in both immediate and week-following opens
• An equal percentage of opt-outs to previous mails
• A 0% reply rate (shocking for us)
• A 14% decrease in click-through (not surprising since there are only two links in this letter)

The text of our mail is included below:

Subject line:

Big Thanks and Best Wishes from Gold Guys Canada!

Text:

Thanks for all the fun

We at Gold Guys just wanted to say thank you for a year marked by some great gold parties and many wonderful people. We thank you for welcoming us into your homes, and for introducing us to your families and friends.

We hope that in the coming year we’ll get to see some of you again.

Best wishes for the New Year

We extend to you all sincere wishes for a 2011 filled with all the things that make life great – laughter, friends, family and growth.

Sincerely,

Lawrence Ruttner, Michel Korman, Barry Korman and Martin Preikschas,
www.GoldGuysCanada.com
www.My-Gold-Party.ca

My take away, due to the significant decrease in opens and replies, is that folks suffer from email fatigue during this time of year AND that they expect bold, call-to-action, value-proposition emails from businesses even if all you want to do is say thanks. In future, if I feel the need to get snuggly with our clients, I’ll save it until I also have something tangible to offer to show our appreciation.

This doesn’t mean 1-800-Headsets wasn’t successful, as customer base, email content and a variety of uncontrollable factors are involved. But Martin’s assessment of the reasons WHY the email campaign wasn’t successful is valid. It’s always important to test, measure important metrics and explore reasons why a result was positive or negative.

I’m not sure what I’m going to do with it quite yet but I managed to get WordPress installed without a hitch – now to make it look pretty….

There’s a lot to unpack in this email.

Merchandising

At first blush, it doesn’t appear this letter is merchandising anything – it’s all text. In fact, it opens by saying it’s not trying to sell you anything. But a careful read reveals something very clever.

“Following a couple of slower growth years, 2010 has been a big year for us, most notably with the release of the OfficeRunner wireless headset, which you’ve undoubtedly heard all about by now, so I won’t take your time going over the benefits it offers.”

I may have ordered a headset a year or so ago, but I certainly don’t keep up with newly released headset products. 1-800-Headsets knows very well I have no clue what the OfficeRunner headset is. This email piques curiosity that there is possibly something groundbreaking that has entered the market that would motivate me to buy again. (I’m unlikely to need another headset for several years, so this is smart tactic).

Value proposition

The email claims 1-800-Headsets.ca is Canada’s number one headset seller. While this would be a stronger proposition if it was backed up by proof or clarified what number one means (number one in service, sales, selection, customer satisfaction), it still communicates a reason to stay loyal to the e-tailer.

Human face

The letter is signed off by the CEO, and includes his personal email address and phone number. This expresses that the company values the individual enough to share the head honcho’s contacts. Though the letter is undoubtedly written by someone else, this is a nice gesture.

Without any calls to action or links, it’s impossible to measure clicks and conversions apart from tracking sales for the OfficeRunner and matching back orders to customer account numbers that received the email. Nevertheless, trying something radically different is a good idea. This email’s purpose was to build awareness and create interest in a featured product, backed up with a value proposition and a warm, we-love-the-customer tone – something the text accomplishes.

Whether this campaign was a success for 1-800-Headsets, I don’t know. Amongst the pile of promotional email this season, customers may have been turned off by the long text or refreshed to find an email that doesn’t oversell. What do you think of this email?

There are a number of reasons why a customer might abandon cart, including but not limited to:

• ”Sticker shock” (reaction to unexpected shipping and taxes)
• Required registration (1 in 4 will abandon unless a guest checkout is offered)
• Security or privacy fears
• Distractions mid-checkout cause customer to close tab or window
• Not ready to buy
• Intent to come back and purchase later (24% of online shoppers, according to Forrester Research)
• Technical issue (form not working, credit card declined, session expiry, slow loading page, unable to bill outside the country, etc)
• Usability issues (forms too complicated, asking for too much information)
• Preferred payment type not available
• Unable to ship to destination

What is a cart recovery email and why should I care?

Not all of the above can be mitigated by conversion improvement efforts like redesigns, usability testing and A/B split testing. Many e-tailers have had success with remarketing programs that send triggered emails to cart abandoners. Cart recovery emails can have upwards of 50% open rates and very high conversion rates since they are relevant and timely. So long as an email is captured in the first step of checkout, retargeting is possible.

Know when to pull the trigger

In 2009, Listrak conducted a study on the triggered email habits of the Internet Retailer 500, and found only 42 that deployed cart abandonment emails. Of these 42, the average time to reach out was 75.1 hours – or 3 days, 3 hours and 6 minutes. Less than 10% sent within the first hour, 17% between 1 and 24 hours, 33% between 24 and 72 hours, and 40% after 72 hours.

Timing your cart recovery email is important. Many email marketing specialists believe best practice is to send within 24 hours of cart abandonment – some even suggest within minutes. However, this advice is no substitute for your own testing. Bronto Software’s Kelly Lorenz recently shared on the Email Marketing Reports blog that one of her clients saw higher conversions and revenue sending 3 days after the abandonment, rather than 24 hours.

Marketing Experiments actually found the best conversion using a series of 3 triggered emails.

Testing is always the best way to know what works best for your business. You may also consider your average days to purchase and visits to purchase (found in your web analytics) to give you ideas on how long your “window” should be.

Crafting your triggered email strategy

Deploy different messages for different segments

While it does make sense to strike while the iron’s hot, if a customer is still in research mode or is intending to purchase later (perhaps when money is available), your triggered message will have little impact. On the other hand, waiting too long leaves room for the customer to shop around or simply forget about items left in the cart intended for later purchase. A good strategy is to deploy different cart recovery emails with different timing for various abandonment situations.

Identify where the abandonment occurred

Typically, the further along the customer is in your funnel, the greater the intent to purchase. Abandonment after entering credit card information could indicate a technical problem with your form. I recommend sending a customer service oriented email offering help with checkout completion (including a customer service number or email contact is helpful) within a few minutes or a few hours.

On the other hand, if a logged in, existing customer (remember, you need an email address) abandons at the shopping cart page, you can assume he or she is either holding their items for later (perhaps to add to their purchase), became distracted, was put off by shipping costs, or is simply not ready to purchase at this time. Testing reminders 72 hours up to 10 days may make more sense here.

Identify why the abandonment occurred

Tagging customers based on what form errors they encountered in checkout forms can help you trigger the most relevant message. If the customer continually received an AVS mismatch error, it could be their card was blocked for security reasons due to several attempts. Most customers don’t understand why, even with triple-checked, correct information, the form continually rejects their input. An email explaining AVS security might prompt the customer to try again in 24 hours, when the “ban” is lifted, or to phone customer service to pay with a different card.

Customers who attempted to enter expired coupon codes may be won back with a personal offer code. (Careful, this may condition customers to always abandon for the prize). Or, you may simply ignore this customer.

Did the customer receive an “out of stock” notice? An email suggesting similar products or offering a “notify me when this item is back in stock” option, deployed as soon as possible would be most appropriate. When the item is shown available on the product page but is unavailable in checkout (this can happen during mega-sales, Black Friday, etc), a small coupon for next purchase is also a nice gesture. There’s little risk of the customer being “coupon conditioned” in this case, as it’s clear it’s a courtesy for the unexpected inconvenience.

If you preserve navigation within checkout (as opposed to using an enclosed checkout which removes navigation), flag customers that click out of checkout to go back to the site. Such customers may be planning to add to their order and checkout later. You may test timing and content (including related product suggestions or other personalized merchandising) to see what works best for these customers.

Just reach out with email?

If you’ve captured a telephone number, a customer service oriented phone call may save the sale. Services like OneCommand can send an automated call to the recipient. But because customers may not appreciate follow up phone calls from online shops, you may want to reserve this for high ticket items where abandonment occurred further in the funnel, or for complex sales that involve customer service (telecommunications, B2B, automotive, etc).

Triggered email content and design

Next post we’ll look at some actual cart recovery emails and deconstruct them to the good and the not so good. Stay tuned!