Oz Analytics

8 Weeks Into Our Beta Launch

It's been sometime since my last blog entry. Apologies, but I have a good reason: I launched my beta website doclens.com. It's now a couple of months on and time to communicate what the experience has been like.

First off, what is doclens?

Our mission is to make data driven narratives by discovering insights from reports and articles.We do this by:

1. Using the online research and opinion published by thought leaders such as consultancies, think tanks and academic experts. We estimate 50,000 – 100,000 items are published annually in our beta target market of analytics and big data people in the financial services industry. We automatically monitor this research as it is published.
2. Our DocLens platform automatically reads, understands and analyses this research and links it to profiles of the people and organisations involved.
3. We then curate the information and publish the best 1% to our website – making it possible for you to keep up with best practices, be the most informed person in your organisation, and make smarter decisions. All in just minutes of your time.

We turn information overload into personal strategic advantage:

Doclens is a lean startup. We have 4 people involved - with only one currently working full-time on the venture. We are also self funded. So resources are incredibly tight, but as the lean movement has taught us, it is amazing what can be achieved.

So here is a summary of the first 8 weeks of Six Degrees of Data as a 'trading entity'.

The launch has been a success. Being a data scientist, I measure success against a set of short-term targets we agreed on prior to launching. These are:

• 100+ subscribers by the end of 2014
• 5,000 article reads within 8 weeks of launch
• Average visit time > 2.5 minutes.

I'm pleased to say that all 3 targets were exceeded.

Whew ...

What else can I say about the experience so far? Well here are some of the highlights, in no particular order of importance or time sequence:

• For 3 of the 8 weeks we conducted A/B tests using 4 social media networks to advertise, using a total of 14 different messages. Twitter has been the outstanding performer for us.
• That said, we have also been actively tweeting and posting about the content being published on doclens each day. Judging by the number of retweets, favourites, mentions and clicks we are getting, this is maybe even more effective than paid advertising. I'm not yet sure how I can test this in practice (given our resource and budget constraints).
• Our customers are of course central to everything we do and some of the organizations they represent are:

• There is no particular geographic concentration in our hundreds of users - at least nothing unusual that I can detect. The USA, Western Europe and Australia feature heavily, as does Scandinavia, New Zealand and Singapore. Our users come from over 80 countries now.
• 95% are data and analytics professionals. Good news as this is are target market.

So what hasn't gone well or smoothly? A few things, that's for sure:

• Social media advertising. LinkedIn, Google and Facebook have not performed well so far. Advertising on Twitter was about 5 times more effective than these other channels. I still don't know why although my 600+ followers on Twitter may be a factor. I need to run some more A/B tests to resolve this.
• Customer engagement. People just aren't commenting on the content we are posting. Perversly, they seem to love the content and we get lots of visits. Our most popular articles have now been read thousands of times each. The silence is puzzling and I will probably survey our customers soon to find out what is going on.

Longer-term (6 month) targets were also set and these are:

• 1,000+ subscribers
• 20,000 article reads
• Average visit time > 5 minutes
• CAC < 40 cents
• Launch of our analyse-your-own-document service.

CAC = Customer Acquisition Cost. In our case, the cost of advertising on social media.

So these are the basic measures we are using. They are useful to have and us help focus on the things that matter.

It has been a slog, with many, many hours expended. And many more to come. But it is fun, we are learning every day and I wouldn't be anywhere else at the moment. I'll let you know if this is still how I feel in about 20 weeks time.

Do you know what is bigger than Big Data?

From episode 10 of my Naked Analyst Channel on YouTube.

I think I do - and it is the ‘appification’ of analytics. What I mean by this is the reduction of a complex analytic activity such as market segmentation, down to a single button on your computer interface. Very much like the Apps on your smartphone, tablet or increasingly your desktop.

That’s what it looks like but the impacts are more profound. That’s because it makes it possible for analytics to be successfully done by people who may not understand how it works, but do understand the ‘why’ and ‘when’ they need to do it.

For example, a marketer in a company can access more sophisticated views of their campaigns without the need of a specialist analyst. Appification extends the range of analytic things that a non-specialist can do.

This appification is made possible because of three things that have emerged in recent years:

1. The rapid increase in the number and sophistication of APIs (Application Programming Interface).
2. The rise of open source analytic platforms like R (http://www.r-project.org). These platforms have created vast libraries of algorithms freely available to anyone who knows how to use the platform.
3. The shear number of people and organisations involved in creating open source analytic platforms like R.

The last enabler needs a little further explanation. R is a free software programming language and software environment for statistical computing and graphics. It contains thousands of packages (10,000?) specializing in topics like econometrics, data mining, spatial analysis, and bio-informatics. Nobody knows how many R users there are, but a reliable estimate (see http://spatial.ly/2013/06/r_activity/) puts it in the millions. Many thousands have helped R develop over the years.  I think that this sort of large-scale self-organising open source effort is beginning to teach the world how to use analytic algorithms.

The above is all supposition, but I can back this up with evidence. Here are 4 examples of algorithm markets - or at least they exhibit varying degrees of ‘algorithm marketness’.

• Dataxu - senses and reacts in real time to changing consumer behavior. Openness is at the technical core of DataXu. The DataXu platform is a flexible technology stack with open APIs that make it easy to integrate and extend functionality.
• algorithms.io - delivers machine learning for streaming data. The Algorithms.io cloud platform makes it easy to use machine learning algorithms to classify streaming data from connected devices. Turn the Internet of Things into the Internet of Action
• Snapanalytx - aim to provide predictive analytics for all and make them more accessible and affordable.
•  Algorithmia - are building a community around state-of-the-art algorithm development, where users can create, share, and build, on other algorithms, and then instantly make them available as a web service.

In conclusion, there is still one issue needing resolution before algorithm markets take-off: How will the world’s business people get data into and out of these algorithm apps? I’m not sure yet, but I think the answer will be more apps. Apps that themselves appify the transformation and loading of data into and out of algorithm apps.

Confused? Well don’t be, like most new and shiny things, what we are talking about is just the next generation of ETL - something business intelligence people like myself have been building for the last 20 or more years.

What’s old is new again? Maybe, but from my perspective the future looks exciting for analytics.

My YouTube channel contains tips from a 25 year veteran of the analytic profession.

What I Learnt Pivoting (so far)

Pivot:    (verb) turn on or as if on a pivot. In a startup it is a strategy for iteratively searching for a repeatable and scalable business model

At christmas we reviewed our strategy based on 6 weeks of feedback from our early adopters. The decision was to pivot and change Bridger from a mobile solution to a desktop one. Another 6 weeks have passed and I have learnt a thing or two that I feel it is worth sharing.

The good news is that there is good news. While pivoting is scary, we all feel that we have a better product now emerging. The consensus is that we are gaining:

1. A product that is more accessible for our users (i.e. it's on the right platform)
2. One that works with a single drag and drop movement
3. Has some very cool new features
4. Presents results more clearly.

We think that we will resume beta testing in about 6 weeks time. So the pivot is looking to have taken less than 3 month's to achieve (assuming no hidden disasters). The mobile version of Bridger took about 18 months to develop from the prototype stage. So the pivot will take only a sixth of that time. Is that fast enough to be considered lean? No idea, it is what it is. Certainly we are continuing to operate on the whiff of an oily rag so any progress measured in weeks is a plus.

At the start of this blog, I said that pivoting is scary because the lean books and websites 'pushing' pivots don't deal with the emotional investment you make in the original concept. It is your baby and as any parent will tell you, criticism of your child is hard to take and even harder to act on.

Although at 6DD we are now all hard at work building the new desktop Bridger, there is also a feeling of of inertia that is still to be fully overcome. Maybe inertia is the wrong word. It's more like regaining momentum after stopping the original strategy and switching to the new.

That's what I've learnt so far and I don't think we are unique in this. So if you are considering a pivot then I recommend that you spend a significant effort minimizing the emotional impact on yourself and your team. It is an intangible element that can suck the life out of your venture if you don't manage it well.

After all, if the empirical evidence tells you to change tack, what other options do you have?

Preparing For Commercial Launch

Following lean principles, we are quickly drawing to a close our beta testing period. At Six Degrees Of Data we're all hoping that the world is ready for a radical new way to use documents and web pages.

At least we won't die not knowing.

A big thank you to all our testers who identified a number of bugs and suggested a number of future features. Your help has been invaluable and as promised we have listened to you.

Bridger version 1.0 is coming later this week with the following requested enhancements:

• Summary texts and the full text original documents now include color-coded hyperlinks to all bridges. The links are to dynamically created profiles. This feature can also be switched on and off as you wish from within the screen.
• Help is now available within most Bridger screens. These include short instructional videos and step-by-step instructions to guide new users.
• We now have a YouTube Channel with a dozen or more videos to help you get the most out of Bridger. More videos are appearing each week.
• Easier navigation through further gesture support. Swipe left and right to move back and forward through most Bridger screens - or use the browser buttons. Your choice.

Finally, and best of all, we have decided to keep Bridger 1.0 completely free for all users. So point your smartphone and tablet browser to www.sixdegreesofdata.com and see how Bridger can transform your documents. It's free to use and there is nothing to install.

So it has been a busy couple of weeks.

Until we release 1.0, there are new videos to check out. Of course, as you use Bridger in the coming days you will see the features mentioned above being introduced. Try them out for yourself and, as always, all feedback is welcome.

Bridger Enhanced Text

One of the neat features of our new Bridger product is the way the text of your original document is enhanced with color-coded links to further information.

As one of our testers recently said, "It's like you turn documents into Wikipedia articles!" - and that's true. But the story doesn't end there as the links are to dynamically created profiles. These profiles are unique to the context of your document. These profiles contain:

• A short text summary created from information outside of the document
• A picture fetched from the web
• In-context occurrences identified from within your document
• Links to places with relevant further information.

Here's a short video that shows you how Bridger enhanced texts and profiles work:

You can also check out our other videos on YouTube and Bridger is now in Public Beta so you can try it out for yourself. It's free to use and there is nothing to install.

Public Beta Release of Bridger

I'm very happy to announce that after a successful closed beta test of Bridger we are now opening the app to anyone that wants to use it.

Our new web site is at www.sixdegreesofdata.com.

What Does Bridger Do?

It reads any document or web page you point it at, tells you what the document is about and automatically creates profiles of all the people, organizations, places, technologies and concepts involved. Bridger profiles are dynamically created using billions of facts from popular internet sites like Wikipedia, Freebase, The New York Times, The Guardian, MusicBrainz, Twitter, and Google.

To Start Using Bridger

Bridger is a smartphone optimised web browser application - not an app downloaded from the Apple or Android stores. Remember, it's free to use, there is nothing to install and access is instant when you register. Android and iOS ready!

If you really want to, then it looks to us like it also works fine on your iPad or desktop running Chrome or Safari. But no guarantees.

Try it and let us know what you think.

Pocket and Bridger

Bridger is intended to be a universal document and web content analyser and we want to have it play nice with as many file types, applications and platforms as is possible. Of course, not all companies are as open as they perhaps could be so we do our best.

We developed the Bridger user interface specifically for smartphones and it works well on Android and iOS iPhones. We also quickly discovered that it does a pretty good job on the iPad as well and we are also working on supporting desktops - although this may only be possible using Safari or Chrome at this stage.

More news on this soon.

This openness is possible for a little startup company like Six Degrees Of Data because we use a very agile set of technologies. These include HTML5, CSS 3.1, jQuery Mobile and others and we've tried to be as open as is possible. But this only takes us so far.

Take off-line reading applications like Pocket, Readability and the like. Both are excellent apps that I use a lot on my desktop browser and smartphone. Like for millions of others, these apps have become embedded in my daily browsing habits. One problem however is that they really only delay the inevitable. If I'm going to get any value out of the documents I've saved, then I'm going have to read them. And this is where things fall down.

I often never get around to reading this growing pile of stuff sitting waiting for me in Pocket and Readability.

The unfortunate reality is that the virtual world mimics my physical world habits. Just like my bedside table is piled high with books and magazines I want to read - so are my virtual repositories: not just Pocket and Readability, but also email, Zinio, Flipboard, Kindle, iBooks, Instapaper and the like.

Now there are precious few things that I am certain of in life but I know for a fact that I will never read everything that I have squirreled away in these apps!

And that's the motivation behind Bridger. It Bridger continues to develop like I plan, then I can use it to extract the value trapped in these documents. But I digress - because there are speed bumps on my road to document nirvana.

For example, Readability uses some url mapping tricks to do what it does and this currently stops us using Bridger to analyse readability documents. A pity, but we will keep persevering until all avenues are exhausted.

Pocket, on the other hand has been easy to integrate into our work habits and the following short video shows you how Bridger can analyse all the documents saved in your Pocket app: