Pear Analytics

It’s been a little over one week since the Pear Analytics Twitter study we posted reached blogs and media outlets all around the world. We were extremely pleased with the outcome and all of the constructive feedback. As mentioned in the study, we will be monitoring Twitter usage and behavior in an ongoing manner, and part of this post will explain what we have planned next.

With the large amount of media attention and exposure our company has received as a result of this study, I would first like to take this opportunity to not only clarify a few things, but ask for your continued input so we can make the next study even more meaningful.

Setting the Record Straight

There were thousands of blog posts, comments, and general commentary about this study over the last week, and I don’t think I’ve ever seen such a wide spectrum of responses, and we welcome all of it.  There were people who loved it, hated it and everything in between.

First, let me say that we are big fans of Twitter.  We use it to share with friends and family, to share our products, to brag about our clients, and in general to share and receive useful information.  Personally, I have found great resources through the folks I follow on Twitter; but I admit, I can’t keep up with all of it and there are an overwhelming amount of tweets that do not interest me.

Second, we’re not telling you how to use Twitter, or that people are using it the wrong way – we simply reported on insights from what we found in an interesting way.

Third, we were not paid to do this study, nor have we made one dime on it.  We also did not spend one dime promoting it.  My friend at Sales By 5 sent one single email to Mashable to see if they would be interested in sharing the report.  They covered the story and other outlets picked up on the conversation.  We did subsequently sent additional notices to media outlets once it began taking off.

Last, I have a personal relationship with Paul Singh, the founder of Philtro.com, and I purposely included his tool in our whitepaper because I think it’s a great product – and it’s FREE; so again, there is obviously no gain in it for us, and all he really gets are some new subscribers to further refine his tool.  We will be partnering on future studies.

Why Are We Doing This

Pear Analytics provides insights to marketers through data analysis. We did other studies on Twitter in the past, as well as other whitepapers on website visitor loss and how marketers can effectively track off-line media.  Every day we meet with clients who are not using Twitter and want to learn more about how to use it, why they should use it, or what other people are using it for. We believe there are usage and behavior insights about Twitter that many are interested in.

Criticisms

As I mentioned before, there was certainly a fair share of constructive feedback about the study, and so I’d like to share what the major ones were and what we plan to do about them moving forward:

Sample Size – many commented that our sample size was too small for the amount of tweets.  Before we started the study, we assumed that there were about 3 million tweets per day in the U.S. alone.  Several of us have math and engineering backgrounds, so we determined the sample size using statistics.  We even checked with some old college profs just to make sure we weren’t completely off track.  The result was that 2,000 tweets would be sufficient.  Moving forward, we will increase the sample size, provided we have sufficient resources.

Subjectivity – many folks emailed us asking what constituted “pointless babble.”  The criteria was this: the tweet did not have an “@”, “RT”, “via” or short URL in the tweet.  If it met this, and did not appear to be useful to a large percentage of your visitors (more than 50%), then we put the tweet into this bucket.  Believe it or not, these became very easy to spot in the public timeline – tweets like “I just saw a raccoon” or “I need to buy some shoes today” fall in this category.  Now, if you are a hunter or the owner of a shoe store, you would argue that those tweets are not irrelevant, right?  Fair enough.  Part two of our study is going to allow real users to vote the tweets as “pointless babble” themselves.  More on this in a bit.

Public Timeline – some folks claimed that by sampling the public timeline, that’s not an accurate representation of the kind of tweets one would receive, and that conceivably, you would only follow relevant people who always say relevant things.  I don’t really believe that, but what we’ll do is let the users decide what is “pointless babble” – that way, it’s from a user who supposedly hand selected the people they want to follow.

Categorization – lots of comments claimed that the categories were vague and subjective. I still feel the categories are fine – but what we can do is sub-categorize on the next round.  For example, on the News category, we could break it out into mainstream, tech, social media, etc.  For Conversational and Pass Along Value, we could add which percentage of those had links.  This keeps the primary categories consistent between reports for comparison purposes.

Moving Forward

As with any analytical study or analysis, there is iteration and refinement.  We feel that for our next report on Twitter usage, the primary focus is to collect data from real users.  To do this, we are partnering with Philtro.com to refine the “pointless babble” category, but will continue to pull random tweets off the public timeline for the remaining categories, but will create sub-categories for further refinement and insight.

We created http://pointlessbabble.pearanalytics.com to show you a live feed of the tweets deemed “pointless babble” by real users and by an advanced algorithm that can detect tweets of similar nature.  Keep in mind this is in beta testing for the next few weeks until we can perfect the process.

Our next Twitter report will be on October 15, 2009.

Thanks again for the support, criticisms, and other feedback.

Ryan Kelly

CEO, Pear Analytics

4 Comments »Tags: twitter

A while back we embarked on a study that evolved after a having a debate in the office as to how people are using and consuming Twitter. Some felt it was their source of news and articles, others felt it was just a bunch of self-promotion with very few folks actually paying attention. But mostly, many people still perceive Twitter as just mindless babble of people telling you what they are doing minute-by-minute; as if you care they are eating a sandwich at the moment.  (See our last post on Twitter: Is Anyone Paying Attention?).

So we took 2,000 tweets from the public timeline (in English and in the US) over a 2-week period from 11:00a to 5:00p (CST) and captured tweets in half-hour increments. Then we categorized them into 6 buckets:

News, Spam, Self-Promotion, Pointless Babble, Conversational and Pass-Along Value.

See our entire whitepaper here.

Results

The results were interesting. As you may have guessed, Pointless Babble won with 40.55% of the total tweets captured; however, Conversational was a very close second at 37.55%, and Pass-Along Value was third (albeit a distant third) at 8.7% of the tweets captured.

Other Results

- 11:30a and Monday’s have the most frequent tweets with Pass-Along Value.  If you want something re-tweeted, try it at 11:30 in the morning, or try it on a Monday.  Maybe folks are feeling generous then.

- Spam seemed pretty consistent all day, everyday, but luckily was second-to-last on frequency.

- News seemed heavier at 2:00p and on Tuesday’s.  Note that we did not collect any tweets during the Michael Jackson news, in order not to skew the results.

- Conversational tweets were high between 2:00 and 4:00p throughout the study, and heaviest on Tuesdays.  Most people are busy on Monday catching up with work, so perhaps it takes until Tuesday to get into the swing of Twittering again.

Conclusion

With the new face of Twitter, it will be interesting to see if they take a heavier role in news, or continue to be a source for people to share their current activities that have little to do with everyone else.  We will be conducting this same study every quarter to identify other trends in usage.

Since Twitter is still loaded with lots of babbling that not many of have time for, you should check out the Twitter filter, Philtro.  These guys can not only help you filter the noise, but will also be allowing you to store the tweets you are most interested in real soon.

See our entire whitepaper here.

Update – August 24, 2009

Please see this post on how we plan to continue analyzing Twitter behavior and usage.  We are looking for your feedback to make the next study even more insightful!

247 Comments »Tags: Social Media, twitter

What exactly is anchor text?

Well, anchor text is a common term we use in SEO to refer to the text links on a webpage.  These links are highly valuable in helping you rank for the search terms you desire.

The concept is fairly simple: make a list of the terms you would like to rank well for in Google or other search engines, and create links to your webpages with that as the linking text.  For example, if I want to rank well for “free SEO analysis”, then I would try to put that as linkable text on as many high authority and trusted websites as I could.  This is why if you were to copy the script for our free customizable widget for the Website Analyzer, you will see a link like this on the widget: a free SEO analysis by Pear Analytics.  The link brings the user to the page I want, and has the proper text in the link.  It is important that the copy on the page has to do with the anchor text as well.  Try to have the words in the link on the page somewhere, preferably in the title tag as well.  If you do any Google AdWords campaigns, you would be familiar with this methodology since this is the same general concept around the Quality Score (the ad has to be relevant to the page it is pointing to).

It works, and we proved it

Last week, my friend Nan Palmero wanted to know how he could rank high for the term “blackberry power user”.  If you know Nan, then you know that he really is the ultimate user of the BB.  At the time, he ranked somewhere on the third page of Google for this search term.  He frequently blogs for some big sites with lots of trust, like blackberrycool.com, mysa.com and others.  I found out that he was linking his name to his blog, so I asked him to go back to all of his old blog posts and change the anchor text from “Nan Palmero” to Nan Palmero, Blackberry Power User.  The next day he was on the first page of Google.

Try it, and let me know how it worked out for you.

2 Comments »Tags: anchor text, SEO

We recently upgraded our logo to represent our new direction towards being a product-focused company. The logo is courtesy of Brandstack, who we happen to co-office with – so it worked out great!  The process was super easy.  Wes and Robert showed me a couple of logos from one of their designers who does great “web 2.0″ brands, and I was hooked.  So I got in touch with “voxsix” on Brandstack and asked him to do several renditions of our existing logo, as we progress into being a product company.  About three days later, he came up with about eight different designs, and we went around and got votes on what the best one was.  The best part of the whole thing was that I only paid $350 for the logo, instead of the usual 10x that at a fancy “design shoppe”.

Why the pear?

Almost everyone I meet asks me “why a pear?” and what the original logo really means. The pear represents the organic, fresh and crisp nature of what we do. There’s no “auto-pilot” here, and everything we do requires interpretation and analysis to develop real insights. That’s why in the original logo, half of the pear is gray, blurry and out of focus – that’s everyone else; whereas the other side (us), is in-color, clear and crisp.

The new logo expands on the same concept, but with a few subtle changes.  First, instead of the straight line down the middle, we now have a sort of this ying-yang effect of bringing balance and equilibrium to your marketing activities.  I also like to think of it like Pear Analytics is seeping over the “dark side”, so as the color side infiltrates, there is more clarity now in the entire marketing equation.  The logo is now in full vector format, since the last version had an image as part of it, so it was difficult to blow it up to anything bigger than a business card.

So that’s our story – what do you think?  If you need a logo upgrade, head over to Brandstack and check out their inventory.  Even if you don’t find something exact, just contact the designer directly and they will usually modify it to exactly your needs.

4 Comments »Tags: brandstack, pear analytics

Have you ever been to a website that takes forever to load?  What do you do?

We’ve taken some past research and developed a way to determine how many visitors you could potentially be losing based on how long your website takes to load from 0-30 seconds.  This was not easy – only a couple of studies have actually been done, and not only are they “aging”, but they have also been controversial and only up to around the first 4 seconds of load time data.  Obviously, there are many factors involved in determining how long you are willing to wait for a page to load, but with tabbed browsing, faster connections speeds, and more, maybe this is why a real study has not been done since 2006.

Here are some key takeaway points from the research we were able to come up with:

- Zona research said in 1999 that you could lose up to 33% of your visitors if you page took more than 8 seconds to load.
- Akamai said in 2006 that you could lose up to 33% of your visitors if your page took more than 4 seconds to load on a broadband connection.
- Tests done at Amazon in 2007 revealed that for every 100ms increase in load time, sales would decrease 1%.
- Tests done at Google in 2006 revealed that going from 10 to 30 results per page increased load time by a mere 0.5 seconds, but resulted in a 20% drop in traffic.

Wow.  A half of a second?  Is that even enough time to take a breath? Yet, when browsing, most people will lose patience and leave your website before they even have time to breathe.   How this relates to e-commerce sites is pretty important. If your website is selling a fairly generic item, your site had better load pretty damn fast or you just lost your sale to some other guy. At Christmas, when every parent is looking for this seasons must have toy, better hope your website loads in under 2 seconds. When a husband forgets his anniversary and is quickly looking for a flower delivery place while the boss isn’t looking, your pictures better not be too big and take forever to load.

So how long does your webpage take to load? Check out Pingdom.com/Tools, and then come back here and approximate your potential visitor loss:

If you prefer to “geek out” and read our entire white paper, you can download it here.  (I will warn you that it does mention words like “mathematical model”, “radioactive first oder decay” and “non-linear regression”.)

10 Comments »Tags: ecommerce, page load times, visitor loss