When we launched the new trunk.ly last week, I promised that we’d highlight some of the new features and why we are the easiest way to save links online.
As we said when we very first launched trunk.ly back in December last year, the nature of bookmarking is changing it’s:
now a rolling social rumble of retweets, likes, favorites, sharing, commenting and general discussion…
Fast forward almost 10 months later and that hasn’t changed. We still firmly believe that the future of bookmarking is not just manually tagging and curating your way around the internet, but rather it’s about hooking in to your natural online behaviours and using automation to collect what interests you and search to quickly find it again.
But the old trunk.ly was still somewhere half way between the two worlds of “traditional” bookmarking and our vision of automation and search. With the new trunk.ly we wanted to achieve several things:
- Make the goal of trunk.ly much clearer to new users.
- Ease the path of new users into the product with a better sign up experience.
- Improve the speed of trunk.ly
- Clean up the UI and set up a new design that will allow us to extend the features without the clutter.
Make the goal of trunk.ly much clearer to new users
Our new landing page clearly outlines the key features in trunk.ly. What’s always been interesting about trunk.ly is that with the exception of occasional PR spikes where someone might write a blog post about us, our new users are primarily driven by word of mouth. This often meant that people hitting the homepage would sign up regardless (because someone had recommended it to them) but that they really didn’t know what we did. They often came because someone tweeted something like “Trunk.ly is awesome!”.
By outlining the new features, we’re clearly introducing users to the benefits of Trunk.ly. This has meant that our “bounce rate” (people hitting the home page who go away) has slightly increased – if I know what you do and it’s not for me, I no longer need to bother signing up to find out. That’s OK – the great news is since the new release the total number of users finding us through search and by referral has significantly increased and although in percentage terms about 10% fewer sign up, in real terms the total number of users that do sign up has doubled and is growing.
Ease the path of new users into Trunk.ly
The clearer home page helps, but the big improvement in people correctly setting up Trunk.ly is mainly due to the new wizard.
This new wizard achieves two things.
- It guides new users through the most important actions that will get them the most benefit from trunk.ly like connecting it to Twitter or Facebook.
- It does so in a way that means that by the time the user arrives at their home page, they typically see their latest links already indexed.
Those of you that have been with us on this journey since day one might remember how woeful this initial experience was. Initially when we launched (especially as we had some scaling problems), new users might wait 4 – 6 hours to see a link. Now when you use trunk.ly for the first time, your links are typically there and waiting for you on the first use.
The previous design lead to all sorts of frustrating issues, including one user who took seven months to understand what trunk.ly was and how it could help him (for him then to become a supportive convert). For every user that persisted for seven months, more would have disappeared. Clearly that experience needed some work. I’m sure we can do better, but we’re really nailing the key issues with this and couldn’t be happier with the take up we’re seeing.
To put some figures around that – there is a 100% increase (or near enough to) in the number of signed up users that complete a key step like connecting Trunk.ly to automatically extract their links from Twitter and Facebook.
Improve the speed of trunk.ly
Lots has been done to make this happen, much of it behind the scenes. One significant change is to the timeline (although it’s still visually similar).
Previously the way this was done meant that something like around 75% of the resources in our database were being used to maintain the timeline. We completely redesigned it and improved it which means that it’s much snappier and we could remove that hungry collection from the database all together.
Other things like the tags were not only primitive in terms of functionality, but the way they were implemented was causing the database to slow down (and users to occasionally see 500 errors). We also re designed these which means that now, they don’t slow the site down and a lot of much needed tag functionality is being implemented (there’ll be some new features soon).
We also make a lot more use of compression and caching to improve the speed. Trunk.ly today is significantly faster than it was previously.
There have been lots of studies which show that there is a strong correlation between the performance of a site and user satisfaction. We haven’t quantified it, but we think this is really important.
Clean up the UI
This is perhaps the most controversial of the changes. While most users love it, we do have some people that strongly prefer the previous version. Judge for yourself:
The previous UI was far too subtle. We were written up by
The Next Web on one occasion who said about the old design:
If I have a criticism, it’s that there’s no main, Twitter-style aggregated feed of all the links shared by people you follow – you have to visit their individual pages.
It’s there on the left in the old design (actually it’s what’s selected in that picture above).
We received emails about the old design that said things like:
Because the only part that interests me, is a tiny tiny sentence… lost in a sea of irrelevant content.
And there was a lot more. We sat down with users and they told us they wanted to be able to search their links. They said that they wanted to be able to manually add links. Both features which are core to Trunk.ly and have been since our launch. But they couldn’t find them.
The design needed to change. The new look introduces a lot more contrast. The use of colour is more subtle, but in turn is intended to draw your eye to important news. Error messages have largely been standardised and work the same way throughout the site.
By moving content to the slide-out panel we could improve the loading speed and predictability of the page layout over the old version (large images sometimes took up the whole page).
The whole site is now much more cross-browser compliant and works well on iPads as well as Internet Explorer, Safari and Opera – not just Chrome and Firefox.
We deliberately went very minamalist and the feedback has on the whole been great. While some users want the old design back, the majority who’ve commented like the cleaner layout, the new fonts and the subtle use of colour. Still, there are a few items we’re reviewing (particularly the edit and delete links) and we’ll continue to tweak it (like we tweak everything) to try and improve it even further.
Where to from here?
There’s a huge road-map ahead of us. In the short-term we’re very focused on polishing this new release. We’ve made literally hundreds of small changes to the site in the last couple of weeks, fixing bugs, tweaking text and improving things in response to feedback. Most days we’d push 5 or more changes to production. There are also few consistent items of feedback we’ll continue to monitor and decide if we want to change the layout a little more in response to these.
We’ve got even more improvements to the tags coming, including some bulk edit features. We’ll implement people search to make it easier to discover and find users to follow. We’re overhauling the emails. We’re launching more connectors (on the immediate list are Disqus, GPlus and StumbleUpon) and we’ll continue to work hard to keep improving trunk.ly and making it the easiest way to save links online.
In summary, this new design has met all our design objectives, it’s cleaner, engaging more users and has the space we need to continue to add and expand on Trunk.ly.
