Haters Are Going To Hate…But They Aren’t Always Right
Our Inital Read: Lots of Complaints!
Our first step was to take a look at the The YouTube blog. We were really surprised at how the boards seemed to be full of complaints and calls for a return to the original layout that users were familiar with. All the while we were left thinking ‘well, actually we kinda like it’.
We thought the complainers might be wrong, and set out to find out what the truth was.
The main rollout of the new YouTube UI went live on the 2nd of December. We used the DataSift platform to quickly set up some streams to capture the real-time public reaction of You Tube users to the new design. (Our platform is so easy to use that someone from our Marketing team actually did most of the work.) The rule we set up for this stream captured any tweet that contained the word “YouTube” AND a word relating to the redesign, such as layout, style, interface, etc. We also added sentiment tagging to see how people generally felt about the redesign. We ran a filter for about 40 hours, and collected over 35,000 tweets from all over the world commenting about the YouTube UI update.
We found the “Haters” were wrong–a lot of people actually liked the design.
While one might have gotten the strong impression from online blogs of widespread public disappointment with the changes, a quick analysis of the tweets that the stream gathered over a 40 hour period told a different story. We found that there was actually a greater expression of positive rather than negative sentiment expressed by Twitter users in reference to the changes. 32.25% of the Tweets captured by our stream expressed positive sentiment, 57.26% of the tweets were neutral and only 29.30% expressed any negative sentiment. 29.30% is still high—but we wondered if there will always be a percentage of users who hate any new change to a UI.
The big surprise – most of the tweeting was being done by influential people!
Although the average Klout score of Twitter users is close to 10, we were fascinated to find out that 53.5% of the individuals tweeting within this stream had a Klout score of 26 or higher. This meant that not only was there a strong expression of positive sentiment but most of the tweets featured in the stream were also being written by highly influential Twitter users. It seems to us that YouTube must definitely be doing something really well if so many highly influential Twitter users were tweeting about the new UI.
What did we learn?
1. Twitter is a great place to get a complete picture of what users think about a new product.
If as YouTube say on their blog, they ‘rely on your feedback to figure out when we’ve gotten it right and when it needs further tweaks’, it’s important that it’s not just the views of a passionate minority of extreme YouTube lovers posting on the company blog, rather than the opinions of the greater YouTube user base, that drive forward future site developments and strategy.
2. The US is important, but so is the UK and Brazil!
The 37,385 tweets that DataSift collected deliver an objective image of the spread of public reaction immediately after launch. A broad data set such as this gives a reliable aggregate view of how people are responding online. This data can be be consumed though our API in real time and funneled through any application (CRM solution, dashboard etc.) or simply recorded and extracted as an excel file. It can also be analyzed by geographical region, gender, degrees of positive and negative sentiment and social influence level, url content as well as the trends within those streams. For instance one insight that we gleamed was that the tweets within this stream came mostly from the US but also the UK, and perhaps surprisingly Brazil.
3. DataSift is a great platform for mining Twitter content.
The DataSift platform is so easy to use were able to set up trackers in just a few minutes that were able to provide very insightful information. With all the noise that comes with the real-time web, extracting the information that will drive productive changes in your company’s communication with it’s customer base requires heavy duty technology. DataSift was architected for enterprises to manage the unstructured data of real-time social media with finesse so they get just the information they need to confidently derive effective strategy from the data.
4. For the daily price of a Starbucks Latte, YouTube could get really interesting insights into their users.
See the stream running live here. This stream costs $4.80 per day to run (not including licensing costs at 0.10 cents for 1000 tweets). Have a product launch coming soon? Set up your own stream on DataSift in less than a minute.