-
UK General Election candidates, who are they really?
Forget reading manifestos, analysing policy impact on your monthly take-home pay or weighing up the pros and cons of entry into the Euro. Who are our political leaders and, more importantly, what do we really think of them?
What started as a conversation at SXSW launched in double-quick time just over a week ago after a flurry of code production and pixel shuffling.
-
Here comes SocialMod
We’ve pushed the first ‘live’ release of Socialmod today. We think it has the potential to become the ‘Basecamp of moderation’.
SocialMod is a hosted moderation service that uses an API to receive items to be moderated (text, images, videos, tweets), processes them and then sends the verdict back to the client’s site.
It’s our first product and the brainchild of our young Rails rock-meister Alex Maccaw. Alex turned up one day at work with a fully functioning prototype and said he’d like our help in designing and selling it, as well as letting him spend some work time on it. Its his company (LeadThinking is the parent company he’s set up) and we’ve taken a small stake. He’s already signed up paying customers, and the service has been enriched through the feedback of several of the world’s leading moderation companies – thank you to Tempero, eModeration and Escalate and everyone who participated in the beta.
Socialmod’s solves a number of problems faced by site owners implementing social software and publishing user generated content.
Firstly, the service brings moderation within reach of ‘the little guy’: the local football club, local community sites, the lower end of market and small businesses. The overhead of creating your own moderation software, or of using a moderation service is often too high for people like this.
But Socialmod also offers moderation companies a number of benefits. Moderation software is – on the whole – pretty rubbish. There are lots of products, each with a different user interface to learn. With notable exceptions, they’re typically quite clunky and often buggy. Moderation companies have to learn how to use each and every system (there could be dozens in use by a single moderation team) and they waste time working round the clunks. Additionally, many moderation tools are bespoke applications that get the job done at launch but are difficult or expensive to upgrade as things change – and things change very rapidly these days as site-owners adopt more and more social functionality. And then there’s the need to integrate with an increasing number of external services – Twitter is the best example. Obviously, there is the scope for moderation companies to re-sell the application.
There are a number of neat features we think will be really useful:
- Twitter moderation (filter on replies/links/re-tweets)
- New profanity settings – option to star profanity rather than block it
- The option to embed moderation into your website using an iframe, so that trusted users can, for example, moderate in situ when signed in
- We’re using an Amazon load-balancer with auto scaling dependent on demand
- There are different plans for manual and automated moderation
- A full audit trail
- A referral and escalation system
- Language filters
- Spam filter
- A profanity filter
- Auto take down (if an item hasn’t been moderated within a specific timeframe – we’ve called this a ‘dead man’s handle’)
- Flag take down (support for reactive moderation)
- An analytics dashboard – a feature we’re planning to expand
- A usage CSV download
Fascinatingly, Alex has also set it up to work with Amazon’s Mechanical Turk Service, allowing site-owners to crowd-source moderation.
Lots of releases in the pipeline and we’ll keep you updated with news. Another great design from Julia.
Related posts:
-
TweetMod – Moderated Twitter Stream
You probably saw the Skittles debacle a while back. They basically turned their entire home page into a Twitter search for the word ’skittles’.
What was quite an innovative idea opened Skittles up to quite some exploitation, people posted literally anything to get on their homepage, regardless of how gratuitous.
Kudos to Skittles for sticking with it – but although many brands want to utilize UGC, they can’t take that sort of risk risk.That’s why we’ve made TweetMod – an extension to Socialmod (discussed previously).
You just need to specify a few Twitter search queries that you want to track and, instead of pulling the feed from Twitter, pull a moderated feed from us.
We’ve also emulated the Twitter API to some extent – so you can use the existing libraries out there for Twitter – just change the endpoint.
If you’re interested in participating in the beta, let us know. We’ll be launching shortly.
-
SocialMod – Moderation Service
Whether you’re building a new website, or want to manage your community more effectively – moderation is a common requirement.
Unfortunately computers haven’t got to the stage where they can recognize libelous/offensive images – so UGC often has to get passed in front of a human moderator to get an accurate verdict.People often build moderation systems from scratch – even though it’s quite a generic problem. Today we’re officially announcing SocialMod – a comprehensive hosted moderation system that prevents you from reinventing the wheel.
The idea is simple; you send any images/videos and text to SocialMod using the API. Once they’ve been moderated, you get a verdict back.
For example, a user uploads an image to your website. Your site then automatically submits it to SocialMod and, once the image has been moderated, you can then display it publicly (if it passed moderation).
Or, if your traffic is even higher, you could just submit an item to SocialMod when it gets ‘flagged’ by your community. Have a look at my previous post on moderation types for more information.So who actually does the moderation? Well, there are three choices:
- You or your team. You can add extra users to your account and they can perform the moderation.
- Your community. SocialMod can be embedded into your site and you can give trusted members of your community access.
- Us. If you choose one of the automated plans, we’ll do all the moderation.
There’s much more to SocialMod, such as spam & profanity filters, referral, audit trails, reports and Twitter integration. We’ll be blogging about some of these in the near future.
If you’re interested in participating in the beta, let us know. We’ll be launching shortly.
-
Moderation types
Whenever companies want to publish user generated content (UGC) they need to look into moderation, which usually consists of the following three types:
Pre-moderation:
Some sites operate on the principle that every piece of UGC should be moderated before it’s displayed on the website. This is the approach Amazon takes for their reviews, for example, and what the BBC use for their message boards for children. It’s usually used for content that isn’t specifically community based, or content that’s potentially dangerous (from a legal perspective). Some message-boards, particular those that concern themselves with topical issues or celebrities, are prone to libel and can be a source of legal anxiety for the organisation that hosts them. Pre-moderation is probably a good idea in those cases. The drawback is the lack of instant gratification to the user (when they don’t see their content displayed immediately) and the negative impact that can have on your comunity.
Post-moderation:
One of the drawbacks of pre-moderation is that it can cut down on the conversations within online communities, especially if the content takes some time to moderate. One solution to this is moderate everything after it has gone ‘live’ – which means you can still have effective community conversations while keeping objectionable content at bay.
There is a legal issue in this method though, as the operator of the site assumes responsibility for any content that appears, even if that content is subsequently removed by moderation. It’s worth noting too that the less time that offending content appears online, the fewer people will see it – and that could impact on liability or reduce an award of damages.Reactive-moderation:
This type of moderation is fairly passive, and relies on the users to flag objectionable content – for you to then review and, if appropriate, remove. The advantages of this approach is that it scales with your community and it’s easier to avoid liability for anything that is defamatory, infringing or otherwise unlawful as long as you provides a process for removing offending content expeditiously upon being made aware of it (generally considered to be 48 hours). The downside of this approach is that abusive content can appear in front of your users, and you’re relying on your community to tell you about it. This might not be acceptable to all organizations, particularly the highly brand conscious.
A recent court case in Italy may have grave consequences for reactive-moderation and the idea it’s a legal safe harbor. If the case is not defeated – the implications could be huge.
SocialMod:
We’ve been developing a moderation platform called SocialMod, which is in a private beta at the moment. We’ll be blogging about this subject more in the future.
Based on Everything in Moderation’s “Four Types of Moderation“.Links:
-
What the fig is burp.fm?
Yeah, good question.
It’s something we’ve been working on as a ‘fun’ side project. I’m not going to try and convince you that it’s a gently mocking critique of Twitter, or of social media in general. I’m not even going to try and sell you on it being a ‘trifle’ we created to cheer everyone up in these challenging times. We created it because we have puerile senses of humour and found it funny – you know, burping and stuff – and wanted to make something playful and plastic enough to fool around with, that would integrate with Twitter and other social services we like.
Okay, we know we’re not going to win any awards with this. We don’t want to get all serious about it. It’s a toy. A plaything. We find burping quite funny – nothing wrong with that. It’s a form of non-verbal social interaction that would have been familiar to the Vikings, or perhaps to cave-men (maybe more the men than the women, although I don’t want to get into trouble for suggesting that women can’t or don’t burp – of course they do and can. Some of my best friends are women burpers). So just try and enjoy it – have a go, let’s collect some burps and see what happens. If you want to send any suggestions about how we might improve it then great. In fact, please help us test it. Yeah.
Many thanks to Al Merry at BBH for talking to us about burps in the first place. Basically, it’s his fault.
-
Metrotwin Recommends
We’ve been using our new Acts As Recommendable plugin on metrotwin.com and it’s been interesting to see how it’s performing in a real-world situation.
Bookmarks (places) are integral to Metrotwin, and a user can associate themselves with a bookmark by ‘Loving it’, saving it to their profile, or by stating they’ve been there.
So there was potentially a lot of information that could be collected about users preferences from their association with bookmarks. And that information could then be used to improve the overall experience, such as recommending bookmarks to people, and showing similar bookmarks – a great example of a practical application to Collective Intelligence.
-
Tutorial: Easy Rails recommendations with acts_as_recommendable
Following up on Alex MacCaw’s post on collaborative filtering. The plugin we recently released acts_as_recommendable allows Rails developers to quickly add some user-driven recommendations of items to their latest great millionaire-making startup. At Made By Many we’ve been developing some great niche social-media Ruby On Rails sites recently with New Bamboo and Headshift. The new edge of social media is in the maths, commenting and rating is so old-school, it’s what you do with that data that counts.
This is going to be a tutorial for simple integration of acts_as_recommendable to recommend your users some books.
-
Recommendations & Collaborative Filtering
Data-based recommendations have really revolutionized marketing and web services, making patterns out of the massive amount of information collected about people in order to give them relevant ads, products, friends and music as well as whole host of other things.
Amazon, for example, tracks my browsing history and buying habits to give me a list of products that I’d hopefully be interested in, and usually their algorithm is spot on. I’ve no doubt that recommendations have contributed greatly to their success.
Likewise, Last.fm indexes my music collection and tracks what I listen to in order to give me recommendations about music I haven’t listened to. Like Amazon, they’re usually give pretty good recommendations.
Delicious is an example of a site that doesn’t make the most of the data it collects. With the amount of sites I’ve bookmarked with their service I’m sure they know what I’m interested in and give me relevant recommendations. Perhaps the data processing power needed is what’s holding them back?
ReadWriteWeb has got a good article on Collaborative Filtering and makes an important point about The Wisdom of Crowds which suggest that “as communities grow, not only does a large (diverse, independent, etc.) community make better decisions than a handful of editors, but the larger a community gets, the better its decisions will be”.
Once you scale past more than a dozen users it soon becomes unpractical to make manual recommendations to people, and computers are pretty good making them themselves. A machine doesn’t need to be ’self aware’ or to actual listen to the music to know what you like. No, what actually usually happens is much more low level – users are grouped based on their listening habits (or whatever else it happens to be), and then users are suggested music based on what other people in their group are listening to.
The fact that it’s so level, means that for a lot of recommendations you don’t need data specific algorithms or code, it’s enough to have a relationship between two entities to get recommendations.
At Made by Many I’ve been working on a plugin for Rails called acts_as_recommendable that makes adding recommendations to your Rails sites a piece of cake! If you have a relationship between books and users, for example, acts_as_recommendable will show you which users are similar and which books a user would probably like to read.
The code is based on the example from the book ‘Programming Collective Intelligence‘ which I really recommend getting. it reveals how collective intelligence can be used in a very practical way with a lot of helpful examples.
Archive for the ‘Labs’ Category
You are currently browsing the archives for the Labs category.
Our latest tweets
- @chrbutler thanks! glad you like it :) RT @chrbutler @madebymany Your blog is great! It's a never-skip for me. 2 days ago
- @chrbutler mentioned our blog in his post on designing for attention, here: http://bit.ly/aTo8eY 2 days ago
- @apophenia in Technology Review says that people will simply carve out a space for privacy as they think best: http://bit.ly/cbbNdy 2 days ago
- More updates...
Most commented
People
Categories
- advertising (9)
- agile (9)
- Clients (29)
- code (1)
- Conferences (41)
- content (6)
- design (17)
- Games (2)
- Great Stuff (13)
- innovation (25)
- interview (3)
- jobs (1)
- Labs (9)
- Made by Many (142)
- mobile (4)
- New connections (2)
- opinion (50)
- play (1)
- publishing (8)
- Social Web (132)
- technology (2)
- Transmedia (1)
Archives
Sites we like
- Aaron Koblin
- Barbarian Group
- BBH Labs
- Big Spaceship
- Big Spaceship – Think
- Bud Caddell – What Consumes Me
- Creative Applications
- David Armano
- Edward Boches
- Faris Yakob
- Ffffound
- Gareth Kay
- Howard Rheingold
- Iain Tait – Crack Unit
- idsgn
- Manuel Lima
- Mike Arauz
- Neil Perkin – Only Dead Fish
- Noah Brier
- Not For Paper
- Paul Davis – Our illustrator
- TED Blog
- Undercurrent
- We Are Social