We are currently experiencing difficulties with the ingestion of content. We are working on a solution and aim to have new content on Viewsflow as soon as possible. Thank you for your patience.
Viewsflow’s performance hasn’t been up to scratch over the last week or so. In our eagerness to deliver better content and rankings for our users we got ahead of ourselves. We gave Viewsflow indigestion and our devs have worked tirelessly to cure Viewsflow of the indigestion.
I’ll outline the changes we made but first some background.
2 weeks ago we boosted (quadrupled) the incoming content. This shone the light onto several issues, which were not apparent at lower content levels. This ranged from standard load issues along with somewhat unexpected issues with our content classifiers. Tossed into the mix was the random failure of one of our servers (something that I could have done without at this particular point in time).
The improvements are on-going as we adjust and tune. We are now in a better position for some the great features we want to implement over the coming months.
Our Daily Briefing is a daily digest of 12 interesting and insightful articles on business or finance.
Each day, the Viewsflow platform scans several thousand content recommendations from smart people on Twitter. These recommendations are then weighted and ranked over the course of the day.
We take a look at the top 15 stories out of several thousand in a given day, and quickly place them in the email that you get in your inbox by 7am GMT.
Some questions people often ask us are:
Why don’t you build those email automatically?
The answer is two-fold. First, the Viewsflow platform is still in an alpha and the various scoring and machine learning systems are still getting trained and optimised. So for the next little while, we need a little bit of manual tweaking – which in turn is a piece of knowledge-engineering that improves the core system.
The second reason is that we want the Daily Briefing to be a human experience.. Computers don’t engage particularly well yet, although we think we know how to resolve this in the next few weeks.
Is the Daily Briefing personalised?
Right now, the Daily Briefing isn’t personalised, although for those of you who register through Viewsflow.com, we can build a personalised view (based on people you want to track or topics you are interested in). We are shortly moving the personalised mailing into an alpha (an alpha of an alpha, is that an alpha squared?) and will be in touch with some of you to test it with us. (If you would be so kind.)
We would love to hear from you
Please leave your suggestions or comments on Daily Briefing below!
We’ve been busy over at Viewsflow and we’d like to share some of the updates to the platform
We already enjoy Viewsflow’s Daily Briefing on our iPhones, Blackberries and Nexus One. We are sure it’ll also look even better on an Apple iPad.
Ipad (courtesy of Gdgt)
So if you like the idea of winning an Apple iPad – and we’d love to see you with one, then become an active subscriber of Viewsflow’s Daily Briefing!
Here is how to enter to win an iPad:
1. Sign-up to our Daily Briefing service.
2. Tweet that you have signed up. (Please make sure you tweet!)
3. Read, and enjoy it.
4. On one day between now and the 14th February, we will randomly select one of the active users to win the iPad.
(If you are already a subscriber, then you don’t need to do anything except read the Briefing as you already do.)
Boring terms follow: Read the rest of this entry »
We implemented a new realtime feed ingestion system over the weekend of Saturday 16 January 2010. Unfortunately, with the realtime ingestion came a flood of new data overwhelming our processing capabilities which caused Viewsflow to go down.
We are in the process of making updates to Viewsflow to better manage the torrent of information from our new realtime ingestion and bringing in some more omph to our processing capabilities.
Viewflow may suffer further intermittent outages while we implement the changes, so please bear with us.
We calculate your influence rank using a range of measures
1. We look at the links that you recommend to your networks and analyse the content and context of those links
2. We look at how well the rest of the world views that content. Did other people believe it had an impact? Did other people tweet that link (or best of all retweet you?)
3. We look at whether you discover / recommend items early, whether you amplify links found by others or whether you are a bandwagon-jumper (nothing wrong with that).
4. We qualify your activity against your normal measure of activity and the levels of activity on a given subject.
We weigh these factors up and come up with an influence score on a per topic basis. From this influence score we build a ranking, and that allows you to say that, ta da, you are the 3rd most influential person on a given topic.
We recalculate these rankings frequently, more than once a day. Our plan is to increase the frequency with which we do these calculations.
Clay wrote compellingly about the notion of ‘algorithmic authority’, the idea that:
it takes in material from multiple sources, which sources themselves are not universally vetted for their trustworthiness, and it combines those sources in a way that doesn’t rely on any human manager to sign off on the results before they are published.
Some form of algorithmic authority or third party-authority measures is increasingly vital online. When networks become to big for any peer connecting to them to have first hand knowledge of the person they are transacting with (or taking information from).
This is why Dun & Bradstreet exists. Or Moody’s. Or S&P. Or the Morngingstar Equity Analyst ratings. It is what made Google’s PageRank work.
However, in the world if digital networks we have had to rely on rough proxies. Perhaps what someone says on Linked In; or more likely who they know on LinkedIn; or by doing a big of Google stalking. In all these cases, the onus was on you–the person doing the due diligence. And the cost was high: one could hardly expect to spend 90 seconds to filter everyone person you come across on a twitter stream.
Another week passes and Viewsflow continues its rapid evolution. This week we are rolling out some asked for features as well as lots of behind the scenes work to lay the foundations for future developments including an iPhone app.
Key developments this week are:
For our alpha users it will be easier to find content and other users who share similar interests.
If you are interested in testing out the service and joining our alpha, please do so here.
You can also follow our twitter feed which will push our interesting business and economics stories each day.
Starting this week, Viewsflow will be more accessible:
We’ve been hard at work over the past weeks to ready Viewsflow for launch later in the year.
Recent developments include:
What this means
For our alpha testers, you can now share content with your friends who aren’t on Viewsflow.
For non-users, you’ll see more Viewsflow URLs spinning across the net. You can also follow our Twitter feed which will push our interesting business and economic stories each day.
If you are yet testing out service and would like to join our alpha, please do so here.