Accuracy of visitor location dramatically increased to 97%

With the addition of a second location database we have effectively reduced the number of occurrences of Location Unknown. A failover component has been introduced.

Until today, when we identified a visitor or user, there was a 17% chance of an inaccurate, outdated, or unknown classification, meaning an accuracy of 83%.

The margin of error has effectively been reduced to 3%, with the addition of a second database to improve upon our geo-location identification technology.

This development pertains to location of website visitors based on IP address assignment.

There is a new CXO in town  … called the Chief Analytics Officer!

Authored by Tomas Higbey

This is just to get you thinking about what’s coming in the next couple of years so that when the CAO shows up in town, you’ll recognize that person and be able to take him or her on.

Just as Big Data is inevitable, so may be the creation of the Chief Analytics Officer or CAO. Companies that are neck-deep in Big Data wrestle with the challenges of multiple data-bases, overwhelming data storage issues and exponential data flows.  In the enterprise sector current trends tell us that individual departments within organizations are creating their own Big Data solutions and analysis. Infighting typically occurs and resources are wasted through the lack of a comprehensive and centralized plan. The smart pioneers that “get it” right from the start, and influence the leaders at the C level will have a major competitive advantage while the competition learns the hard way, wasting internal political capital in the process.

The CAO can create the Big Picture design on Big Data for the entire organization, find which tools work, what projects have performed best, which data sources can be trusted, and measure both the entire ROI for the company as well as rank and rate individual projects.

The reality is that Big Data is showing up in nearly every industry and in every department; in productivity, innovation, competition, and market knowledge. In other words Big Data is being used in Marketing, Operations, Human Resources, Product Development and Finance. There is no place within the organization where it isn’t being used.

Demand for Data Scientists and certified professionals will also increase. A CAO can create the Big Strategy to mitigate the foreseeable challenges and simultaneously keep Big Data’s importance in the minds of the entire leadership in the organization.  No doubt the elite companies are already beginning the planning and even implementation stages for the unavoidable future of Big Data.

But what does this really mean? We start with the issues of Big Data.  Big Data at the core is about the “art of the possible”. For the first time, even small companies have a fishing line, a fishing net and various tools to dive into the information ocean and extract all types of specific “species” that are most important for the business.  By recognizing patterns that would otherwise remain hidden, correlations can be identified. Think: identifying successful patterns.

Data scientists have discovered that orange painted cars are in better shape than other cars.  The “why” is not clear but that doesn’t really matter. When seeing it from the Big Data perspective, we can use this as a guide for buying used cars. This is a balancing between “intuition” and instinct (which are often wrong) the potential risk of which can be curtailed and mitigated by Big Data.

Here are the key reasons why the CAO might be critical for Big Data initiatives:

• Revealing hidden strategies over-looked by departments
• Educate organization (not just C level) about BD’s critical utility
• Organized best talent and practices while attracting the best people from a scarce hiring pool
• Data trumps intuition, meaning a new leadership philosophy is being born, one which must be evangelized within the organization
• Establish standardization of reports / analytic models across the organization
• Establish what is and isn’t possible from the Big Data perspective
• Leverage Big data as a key asset / resource for everyone

To repeat: this is just to get you thinking about what is coming in the next couple of years so that when the CAO shows up in town, you’ll recognize him and be able to take him on!

Q: How do I get my app approved for the Apple App store.
A: Easy, just follow the guidelines.

Ok, so obviously it’s harder than it sounds. There are more than one hundred official reasons why your app may be rejected. On the other hand, it’s very similar to Google’s PageRank guidelines, which have been very transparent in the past: just follow the rules and everything will turn out fine.

1. Follow the guidelines
2. Submit your app for review

Getting your app approved can be tricky. Opentracker can assist you in this process. We can show you:

1. what device Apple used to check your app, for example, an iphone or an ipad
2. what OS version Apple used to test your app
3. how many and which specific events (swipe, button pressed, etc.) Apple generated to test your app

This information can help you gain insight into the app approval process, specifically regarding how your app is tested, increasing your chances of success.

The process can be frustrating and time-consuming, especially as a lot of the guidelines are meta-points that you should probably take into account in the initial stages of conceiving your app. If you would like to see the guidelines yourself, you will first need to register yourself as an Apple Developer.

Here are a few sample causes for rejection taken from the Functionality section:

• Apps that crash will be rejected
• Apps that exhibit bugs will be rejected
• Apps that browse the web must use the iOS WebKit framework and WebKit Javascript
• Location data can only be used when directly relevant to the features and services provided by the app to the user or to support approved advertising uses
• iPhone apps must also run on iPad without modification, at iPhone resolution, and at 2X iPhone 3GS resolution

Regarding this last point, we have a case study. In this example, the app was submitted for review and rejected.
Here is the event-stream of the engineer at Apple who reviewed the app:

In our case study, the app maker submitted an app which was built for iphone and was rejected because it was not compatible for ipad.

By reviewing the clickstream, the app-maker is able to see that Apple tested the app only once, and only on an ipad. It is also possible to see that the platform/ OS was 5.1.1 and that the engineer generated an event stream with 10 events, including 6 buttons pressed.

So, our app-maker can now benefit by getting an idea of how testing is done; how much time is used, how long before testing occurs after the notification email from apple, and crucially, which specific functionality is tested.

And most importantly, using this insight, the app-maker can test on the correct device and OS in order to ensure everything is working and that the apple tester’s events can be duplicated, for trouble-shooting purposes.

So why build and program more software anyway?

There is a tremendous amount of programming going on in the world. As we saw at The Next Web conference  last week – there is no shortage of inspiration and new ideas. App production is growing exponentially. The challenge is to ensure that the technology we are building actually gets used.

In our case we built mobile app analytics because we identified a problem which we believe we are uniquely positioned to solve.
There are now more than a million apps available, taking andriod and ios into account, with thousands being added on a daily basis.

Q: What is the problem?
A: The problem is that there is a large disconnect between most app-makers and app users. App builders build an app and throw it into the world to sink or swim. A great many apps are downloaded and launched just one time.

We think that the likelihood of success for apps greatly increases when app makers take user behavior into account. So we have taken the first step towards solving that problem by building a real-time reporting interface which shows all user events taking place in an application as they happen. Obviously the reports can then be slice & diced and cut up to suit your needs, but the main point has been stated; usability feedback is crucial to success, and this blog post is officially over.

When we say that we have built Mobile App Tracking – that can raise some questions. Definitions are needed.

Q: What is mobile app tracking?
A: Mobile app tracking is combined data from native and web app usage on mobile devices. This means tracking and measurement of app usage. If you build, publish, advertise through or sell apps, you have users, it is important to know how these users make use of your app, and if they are loyal users, meaning retention.

What this amounts to is usability testing. Your app users effectively become testers. Individual profiles mean that you can inspect both one-night-stand installs and loyal users in order to determine what is successful and what is not.

For ten years, we have pioneered real-time website tracking, meaning we show what website users are doing while they do it. Using our knowledge of how to both measure and display real-time data, we now provide this service for apps.

Glossary:

• Native app: An application developed especially for use on mobile devices, written in a native programing language like Java or Object-C and can be sold throuh app stores like Apple’s Appstore or Google’s App Market.
• Web app: A site built using standard internet languages. In other words: people accessing websites through mobile devices.

In our next post we will Define and Discuss the change in terminology we have made from website analytics (clickstream tracking and pageviews) to App Analytics, which is the measurement of events, which we display in Eventstreams.