Lucid Insights

Welcome to Lucidel 2.0 Part IV

Hello everyone! Today’s post will be a brief one about the labs section of Lucidel 2.0. 

The labs section was created to allow you to more easily know if you’re hitting the goals you’re setting for yourself or if changes you’ve made to your site are impacting visitor traffic. In other words, it’s Lucidel’s Before/After tool.

All you need to do is select any day at least 7 days before the current date and hit compare. 

Note that the dashboard includes the same functionality (seen above) in the “Compare To” buttons. Here, you can select either last quarter or last month as a point of comparison. The labs section, however, allows you total control over which date you would like to select. This means that if you know you launched a new version of your site on July 15th, you can choose July 15th and see what’s been going on since. 

Since the labs section is doing calculations in real time, do not be surprised if the page takes as long as a minute or two to properly load. Statistics is tiring work!  

In the end, the results will be worth the quick wait and the labs section displays a “dashboard-esque” display of all the changes that have happened broken down by the Acquire, Engage, and Convert stages (as before) and informs you of which segments have contributed the most to the changes. 

That just about wraps up the Labs section. As always, questions and comments are welcome and if you would like to see any new features implemented, please reach out to me at meylakh@lucidel.com! 

Welcome to Lucidel 2.0 Part III

Hello everyone! Today’s post will be a brief one about the log section of Lucidel 2.0. 

The log section is designed to organize your website’s history for you, keeping track of the biggest changes that take place every week and allowing you to easily retrieve information. Each week consists of a series of insights (very similar to the dashboard) which include the date as well as the insight. 

At the far right of each insight, you will also find two buttons, a graph icon and a question mark. Clicking on the graphs icon shows you the graph that matches the insight.

In the example below, I clicked on “Firefox users are more engaged than others. They view 18.7% more pages and spend 15.4% more time on site.” The result was the graph below which plotted both Firefox (blue) and non-Firefox users (red) on charts measuring two different engagement metrics, minutes and page views per visit. 

The question mark icon tells you more about the insights being presented and gives the numbers behind the information. 

On the graphs that are designed to display spikes, you will see the area of expected movement highlighted in yellow and the line for the actual movement in darker orange. As we can see below, towards the far right of the graph, the bounce rate exceeds expectations and spikes. 

That wraps up the Logs section. As always, questions and comments are welcome and if you would like to see any new features implemented, please reach out to me at meylakh@lucidel.com! 

Tweet

Welcome to Lucidel 2.0 Part II

Welcome back readers! Today we will keep moving through Lucidel 2.0 but focus instead on the Graphs section. 

Just like you’d expect, the Graphs section allows you tremendous flexibility in the type of graph you would like to see as well as the information you want it to display.

In the top section, you can select whichever segment you like and see its impact on any metric over the last 90 days. In order to organize the information, we have classified the metrics by which stage of customer attraction they refer to. Lucidel uses three stages of customer goals: acquiring (increasing raw visitor counts), engaging (improving the duration and quality of experience visitors have once on site), and converting (getting visitors to, for example, purchase your product). We have also added our own loyalty metrics to the engage section, including longevity (expected days as a user), addiction (the number of visits per week), and loyalty (expected lifetime visits) so that you can more accurately pinpoint which visitors have the highest ROI and thus have a better estimate of revenues. 

It is important to notice at this point that for metrics that are US centric (such as education and political party), the percentages that you see should add up to 100% since we are taking them as percentages of your visitors who are from the US - no need to worry about the fact that high school has no meaning in France and that Republicans do not exist in the UK! 

Since you will want to see how you’re stacking up, we have also included comparison functionality. If you select the category that you think best describes your website, we can tell you how other websites in your category are doing (ie what is the average bounce rate for tech websites) using industry accepted categorization. On top of this, you can compare your numbers to the population at large, using the US census bureau’s population estimates. 

In contrast to the top section which always uses 90 days, the bottom section allows you to choose the time frame that most interests you and a graph like the one below will be shown. Not only will it display the actual numbers (orange/yellow) but it will also display the results of a complex algorithm (red) that identifies weekly and yearly cycles, and makes a projection for the direction of the trend. At the bottom of the graph, you will also see a message informing you of whether or not there is a statistically significant trend taking place and if there is, what the rate of change is. Here, we can see that there is an upward trend and unfortunately bounce rate for lower middle class visitors is increasing at 90% per quarter! 

Welcome to Lucidel 2.0 Part I

In the next week or two, we will be releasing the updated version of our Lucidel beta software. It will be sleeker, more flexible, and bring with it several new features designed to improve your ability to understand your visitors. 

In the build up to its release, I will be re-introducing you to our product and showing you its new features, one page at a time so that you can hit the ground running as soon as you install. 

The first page that you will see is the Dashboard. The Dashboard is the home base of your Lucidel analytics. 

Along the left you will see the menu which guides the entire structure of Lucidel. As you can see, the Dashboard is joined by the Graphs, Log, and Labs sections (see future blog posts for more!). Unlike the other sections, however, the Dashboard is further broken into “Acquire”, “Engage”, and “Convert” - the three stages of interaction with your customers. The “Acquire” section will tell you how many visitors you have (as well as which segments are growing) and how many of them are bouncing, the ”Engage” section will tell you how they interact with your site both in terms of time and the number of page views, and, lastly, the “Convert” section will tell you how your site is doing in terms of your goals - this section need not be limited to financial performance but can also include other goals outlined in Google Analytics such as “number of pdfs download” or “number of people who joined our mailing list”. 

It is important to note that you can also utilize the filter to only show you results within a certain demographic - for example, let’s say you wanted to see how you were acquiring, engaging, and converting lower middle class families - simply, click “Select Segment” and choose the appropriate segment and all the results that you see will be related to the segment you have selected! Isn’t that awesome? 

I also want to draw your attention to the two small icons that appear at the far right of certain insights (seen below). As you can tell, when you hover over the question mark, you will be able to see the number that is behind the insight. In the example below, you now know that the visits are declining at 45% per quarter. 

The other icon below the question mark is a tiny chart. If you click it, you will immediately get a visualization of the insight that was just presented. In the example below, we can tell that bounce rate has been steadily increasing over time. 

The last key portion of the dashboard is the right nav bar which contains two things: your status report and links to some of the articles from the very blog you are reading right now! This status report will tell you what additional steps you should take in order to more fully take advantage of Lucidel’s resources. In the example, below, we can see that the user has yet to set up goals! 

Well that just about sums it up folks. If anything that you’ve seen or read seems unclear OR if you have a fantastic idea for a feature that we have yet to implement, please email me at meylakh@lucidel.com. We love getting feedback and hearing about ways we’ve made you happy and ways we can make you happier!

Look for more blog posts coming up next week! 

Knowing Your Unknowns

Today’s blog post will be a short one about being cognizant of information that you don’t know. As analysts, our tool of choice is information and we often get so excited about the data that we do have that we forget to adjust for the data that we don’t. 

Here’s a quick example. One of the websites that I work with has traffic from a number of sources, the largest of which are Google and Wikipedia. Obviously, demographically speaking, the webmasters and I were completely in the dark on whether or not these visitors were male/female, professionals/students, rich/poor. We did, however, have a mailing list and that mailing list asked for some of this basic information, leading us to build up a picture of the typical user. 

In looking at their mailing list’s break down, however, we all walked away thinking that males made up a majority of their traffic and that college or maybe graduate-level educated visitors would prove to be the most loyal. After applying demographic information to their visitors, the reality was that women were actually coming in much more frequently than men (and contributing a higher lifetime value) and that high school students were in fact the most loyal. The fact that we knew demographic information for the mailing list did not justify making assumptions about their greater traffic because the biggest sources for them were a mystery. That was a really cool thing to realize! 

So what’s the lesson? It’s easy for all of us to focus on information that we have but when we overemphasize our “knowns” to the point of ignoring the “known unknowns” (for example acting on an unverified assumption), we can land ourselves in a bit of trouble. 

Had experiences with “unknowns”? Found clever ways to dodge them? Questions and comments are always welcome! 

Understanding the Nuances of Analytics Data

In today’s blog post, I decided that I would go off the beaten path and discuss the natural (current) limitations and realities of today’s web analytics. Make no mistake, analytics is still definitely an incredibly powerful and necessary tool but to be better we must understand its weaknesses. 

Before we begin, I want to remind my readers that analytics software tracks cookies, not people (see “Lucidel Gets Demographic Data” for a brief explanation of cookies) and that is where most of the problems come from. 

1) People using multiple browsers  

Browsers pose the first major problem for web analysts. It is common for computer users to have more than one browser installed on their computer, be it Internet Explorer, Chrome, Firefox, Safari, or Opera.

This poses a problem since because each browser’s cookies can’t communicate with each other and thus are treated differently (as different users), only disjointed data can be gathered. 

As more and more people buy cell phones AND tablets AND computers, their browsing habits can be more and more fragmented, making it more difficult to build a complete demographic profile and skewing visit counts for websites (think about the times you’ve opened a website in one browser, then another, and then opened it on your phone or tablet on the way home). 

2) Computers 

Sharing is caring (or so they say) and computers are no exception. Although this is decreasingly frequent, families sometimes still have at least one computer which they all share. This makes things tricky because amazon.com may think they only have one visitor since there’s only one cookie but really they have four - mom, dad, uncle Steve, and grandma Ann. 

3) Cookie Deletion 

By their nature, cookies are deletable files. Every time that you wipe all of your personal settings, you delete the cookies that were placed on your computer, possibly taking analysts back to square one on building your profile and once again skewing statistics on visits, visitors, etc. If you visited amazon yesterday morning, deleted your cookies last night, and re-visited this morning, amazon sees two different visitors when really there should be one. 

Quick note: you might think that needing to log in to websites might alleviate this problem but in reality logging in only alerts the given site’s internal servers that you, yeah you, have returned. It will not tell google analytics software who are still blind about the fact that you’ve returned. 

^ Not these cookies! Although they do look good :) 

4) Disabled Cookies 

In the same vein as number (3), disabled cookies block our ability to build a profile. My feeling on the matter is that people are actually disabling cookies  (via their settings) less frequently but are increasingly using equivalents of Chrome’s Incognito mode which amounts to the same cookie-less situation. 

5) Disabled Javascript

The practice of disabling Javascript is also increasingly rare because it blocks even the most basic web functionality like drop-down menus as well as the most fundamental of tools (like Google Analytics) from being able to run on a site. Since most analytics uses Javascript, this practice can cause you to “go dark” as far as any analytics or ad companies are concerned. The good news is that according to a 2010 study, found here, the maximum percentage of your web traffic that you should be expecting (assuming your slice of web traffic is not behaving exceptionally differently) to have Javascript disabled is around 2% (United States average) and the minimum can be a quarter of a percent (Brazil average). Part of the reason why Javascript is so rarely disabled is the obvious fact that it adds elegance and makes many of the new, advanced features of the web possible.  

Before I leave you all, I wanted to inject some optimism. First of all, while all analytics tools suffer from these limitations, we are all working to improve our ability to accurately capture data. Secondly, many of these problems (family computers, shut down javascript, etc) are becoming less widespread over time as we all become more tech savvy and as more sophisticated websites are built. Lastly, bear in mind that 10, 20 years ago, absolutely no information of this sort existed. It is incredible that there now exist software that can tell you about your visitors, where they come from, and what they do on your website. Even if that information isn’t perfect all the time, it sure beats nothing and can go a long way to improving your site’s ROI. And as the baby below thinks, that ain’t half bad. 

Questions, thoughts, comments? We always love to hear from you! 

How Lucidel Gets Demographic Data

One of the most common sentiments I have witnessed concerning our product is surprise and curiosity concerning our ability to gather demographic data. “This is a really cool product … but …. HOW do you know that my customer is a man or a woman again?”

Today I’d like to satisfy your curiosity by providing a  brief overview of how it is that Lucidel gathers demographic data. 

Anonymized information exists about all of us on the internet. There is an entire industry that is growing based on the collection and sale of data in an effort to allow businesses to more precisely target their customer base. These advertising companies always use several sources of information in order to build accurate pictures of the consumers. This is why Dad is always getting ads for those shiny new golf clubs and your Mom is always getting ads for new hair products.

The first thing that it is important to clarify is that a demographic profile is always built around a computer, not a person, and never has any personally identifiable information - in other words, ad companies don’t know that it is John Doe looking at the website. They only know that it is a upperclass, middle aged male. John Doe will look the same to an ad company as Al Doe, his twin brother, and as Carl Johnson the guy from 5 states away with the same demographics. 

The way the profile gets built is incredibly complicated with new methods and technologies appearing every year but in essence it follows the following pattern:

• You go to your favorite site X to book tickets for Y. This requires you to have an account or fill out information (such as gender, age, etc). Often times either an advertising company will have partnered with site X to have a couple of lines of code placed on their website.
• The advertising company will use cookies from different websites to error check and to build a more accurate profile. Cookies are small files that are saved to your computer that include saved preferences, passwords, and more. Whenever you open up a browser and find yourself already logged into Gmail, you have a cookie to thank! I should also mention that all analytics software (including Google Analytics) rely on cookies. 
• Over time, advertising companies will not only cross reference one website’s cookies with another’s but they will also use offline data to build a richer profile (using your IP to find out the street you live on as a confirmation of your average income). 
• As we move forward, ad companies are getting increasingly more clever about ways to build profiles (ie new abilities to understand if there are multiple people using one machine, etc) to ensure that the data that is being sold is accurate. 

Loyalty Part I

Loyalty. The quality that we have always valued in our friends, occasionally valued in our general marketing strategy, and never valued in our websites. Why do current analytics tools fall so short when it comes to loyalty as a metric? 

In this post, I want to spend some time talking about where the idea of loyalty comes from, what its worth is, and what questions we should be asking to get a better estimation of its value for our websites. 

As a company (regardless of whether or not we are exclusively e-commerce), we do and should attach importance to whether customers like our product. The most easy-to-recognize symptom that people like the product you offer is whether or not they come back after they use it.  Since customer acquisition requires such steep financial overhead and customer retention is easier and therefore cheaper, it is important for us to understand how we are measuring loyalty so that in turn we can know whether or not we’re getting the most out of our clients. 

In looking at concrete examples, if users tend to come back repeatedly, there is a higher chance that you will have more conversions. This logic is as simple as the fact that the more times users shop for item ‘x’ on your website, the more likely it is that they will eventually buy it. Even outside of e-commerce, users who, say, make a habit of perusing the articles on your website will be more likely to download your latest white-paper or commenting on your blog post (both can also constitute important conversions!). 

Having framed loyalty as part of a larger narrative of making sure your website is successful, we should think about which questions are the most important in deciding loyalty. One of the failings of Google Analytics is it would have you believe that the answer is encapsulated entirely by frequency and recency. While both are certainly interesting metrics to have, they do not answer some important questions.

For example, even if I know 90% of my visitors have only seen 1 page, what does that really mean? If 85% have had 5 visits, does that mean that all 5 visits took place in the past month or in the past 5 years? If 70% had last visited my website yesterday, how many of them visited the webpage multiple times? It seems that some sort of metric encapsulating time horizon and number of visits per day is important and since we don’t know that, we are led to the obvious question - 

Do we really know how loyal our visitors are? 

Thoughts? Comments? Let me know! 

The Importance Of Setting Goals

Web analytics is an incredibly powerful tool and so every day, more and more companies are jumping on the band wagon and installing their choice of analytics software. The singular obsession with features and data for their own sake without any sense of purpose or direction is worrisome. For me at least, it reflects a misunderstanding of the core of analytics - measuring concrete goals. 

I recently had a conversation with a small restaurant owner who candidly told me that he wasn’t very interested in getting web analytics because rather than having his managers’ faces stuck in mountains of data, he wanted them to go after the goal of talking to every person who walks through the door of his 50 seat restaurant. 

While I disagree in that I think web analytics can serve a purpose even for very small businesses, what I really liked and respected about his response was the emphasis he placed on his objectives. Implicit in his statement is the idea that data must be used as a measurement rather than merely a cause for blind bragging rights. The truly astonishing thing is how rare that mentality is. It is shocking how often analytics users will claim “I have x number of visitors every month! Isn’t that fantastic?” 

The short answer is “maybe?”. The fact of the matter is if you don’t know where you are month to month relative to other points in time and don’t set goals for improvement, it will always be impossible to tell if you’re doing well or actually failing quietly. 

But Meilakh, you might say - what other metrics matter as much as how many people come to see my website? What other goals can I set?

From my perspective, the number of goals you can set is practically endless. Heck, you can even keep your visitors/month metric and set goals for the increase you want to see over the course of a 6, 12, or 18 month period. Moving away from basic visitor numbers, you can also focus on the actions your visitors take - how many pdfs were downloaded? how many videos were viewed? Did x number of people sign up for the mailing list? These actions are often more indicative of users who will have a more substantial contribution for your company than visitors who briefly visit and then leave (even if they make your visitors/month metric look really good!). 

The last thought I want to leave you with will be the topic of future posts - the role of segmentation in creating goals. Just as a thought, wouldn’t it be cool if you could create goals for different demographics within your population - be it gender, age, education level, geography, or something entirely different? 

Just a thought.

Have success stories with setting goals? Failures? Thoughts and comments are welcome! 

Best Practices for Facebook Campaigns

Last time, we touched on the importance of social media in the grand scheme of promoting your product, especially in terms of generating a meaningful dialogue with your customers. Today, I’d like to move from Twitter and discuss Facebook’s role in interacting with your current and potential clientele. 

One obvious advantage of Facebook is the added flexibility in both the size and format of the information you’d like to post. From longer status updates to videos to photos, Facebook has a framework for virtually any kind of message. Today, I’d like to take some time to highlight 7 of the coolest practices that companies with outstanding Facebook accounts have: 

1) Using Facebook as a support page.

    It was both pleasantly surprising and impressive to see how seriously some companies took monitoring comments on their Facebook pages. Both complaints and questions were promptly responded to and customers were given information either on how to fix their problems or where to get the appropriate help. Since Facebook is such a readily accessible gathering point, it made the process of getting support even easier and gave people a new forum in which to be assisted. 

2) Making things more social.

    The unique thing about Facebook is that unlike LinkedIn or Twitter, it operates almost entirely in the “off-work” world. Companies are clearly much less reluctant to have non-business related posts, often including statuses for holidays, posting funny topical memes, and even occasionally having contests where they partner with other companies (such as name that song for 6 free months of Spotify). I honestly liked seeing the relaxed side of companies and so do millions of their followers. 

3) Leveraging Facebook’s poll/survey infrastructure.

This one’s fairly straight-forward. Facebook does a wonderful job in allowing companies to create polls and many companies ended up utilizing them either with business-related questions or just fun questions that would attract more visitors and more “likes”.

4) Giving real glimpses into the company.

    Companies approached this objective from any number of angles. Most used their “About” pages as a place to discuss company histories, mission statements, and general goals. Taking it a step forward, many companies included a page that discusses the companies values in a way that they hope will reflect postiviely on their brand - for example some discussed their staff’s volunteer hours while others cited how their work supported causes such as same-sex relationship equality. Lastly, companies often linked their Facebook accounts to Youtube channels which featured anything from tutorial videos on how to use the product to mini features/interviews with the staff (and one company even had a video that showed “A day in the life of …”). 

5) Providing information on events.

    Not only did companies do their basic due diligence and announce upcoming events that concerned them and their industry but they also took the opportunity to link to companies and events where they provided technological support or were sponsors.

6) Publicizing customer testimonials and support.

    Since Facebook already has such a vast network of users, some companies chose to actually connect testimonials to the respective users’ Facebook pages. Thus, when they presented customer feedback, they could provide pictures of the users, making the endorsement appear more legitimate and allowing future customers to more closely connect with the company’s current client base.

7) Attracting more likes.

    On the surface, Facebook likes offer a basic reflection of the company’s popularity. Companies, however, cleverly moved past organically generated “likes”. Instead, they often used them as keys to accessing hidden promotions and features of the company “artificially” pumping their fan base up while also providing a new forum for sales/promotions which attracted new clients. 

Please respond with any ideas or suggestions you might have for how your company has used Facebook successfully!