Fredric Newberg February 23rd, 2010 - No Comments »
We have added support for dashboard channel tracking in our latest release. Instructions for how to instrument your application to track the dashboard channels is available here. The RESTful API reference page has been updated with the two new API calls (‘message sent’ and ‘message response’) that have been added to enable this tracking.
The Kontagent Team
Jeff Tseng February 18th, 2010 - No Comments »
Release 1.16 is out.
We are very excited to announce the a new feature called Custom Dashboard that allows you to create your own customized dashboard pages from any of the time series data that is used in any of our graphs. This feature allows you to add multiple data series to a graph, to mix and match various time series data by adding, subtracting, multiple and dividing data sources from one another. It is even possible to offset a data source by a certain number of days and either graph or combine it with other data source. Here are some examples of the types of things you can do with custom charts.
Subtype filtering – With the subtype A/B testing framework we currently have in place, it is not possible, for example, to see how certain subtypes for invites perform over time. With custom charts, it is possible to graph the number of messages sent, or received of a certain subtype over time. To do this, follow these steps:
- Add a data series and name it “Invite Subtype”
- Select “Invites Sent” as the data source
- Click the “Add Filter” button
- Select the subtype filters
- Save the chart
Now you have a graph that displays an event by subtype filter over time.
Installs Growth over the last 7 Days – Often time, you want to see how a given metric performs over a your historical baseline. Given the very weekly cyclic nature of Facebook applications, one of the easiest ways to do this is to compare the same day last week. With custom charts, you can do this very easily. You could graph the percentage daily uniques growth vs. 7 days ago. To do this with the custom charts, you would:
- Add a data series and name it “Install Growth vs. 7 Days Ago”
- Select “Install” as first data source
- Add another Data Source
- Set the operation to “Divide by”
- Select “Installs” as the second data source
- Set the offset of the second data source to 7 days
- Save the chart
Now the resulting graph will show how your application installs has grown over the last 7 days!
Total Invites and Stream Installs – Another example is if you want to create a custom metrics. This could be something as simple as the total number of daily installs from Invites and Streams. To create this custom chart:
- Add a data series and name it “Total Installs from Invites and Streams”
- Select Invite Installs as the first data source
- Add another Data Source
- Select “Adding” as the operation
- Select “Stream Installs” as the second data source
- Save the chart
Voila! Now you have a chart that shows the total installs from invites and streams over time.
We believe custom charts is an extremely powerful tool and we will be continuing to add features and functionality to it to make it even more flexible and useful.
The Kontagent Team
Jeff Tseng January 22nd, 2010 - No Comments »
Release 1.15 has just been deployed. This is a minor release as we’re preparing for some fairly major feature additions over the next couple of versions. Here is a list of the changes to Release 1.15:
- Daily Retention – For Premium and Enterprise customers, we have added daily retention in addition to weekly and monthly retention. Daily retention can be found under the Retention tab. Please note that for most customers, daily retention data will only be available for today forward, we did not go and back process the daily retention data. If you would like your daily retention back processed, please email us at support@kontagent.com . We have also fixed a number of bugs that prevented data from being displayed with the retention graphs.
- “Stickiness” – The “stickiness” metrics is a metric that has grown in popularity over the last months. The “stickiness” metric is DAU/rolling 30-day MAU. We have added this to the retention tab. Please note that MAU and “stickiness” data will only be available from today forward.
- Last Updated Times for Traffic Source and Insights – Almost all the time series data on Kontagent is updated hourly, but some pages like the Traffic Source and Insights are updated daily. We have added a “Last Updated” field to the Traffic Source and Insights pages so you know when the daily data was updated.
As always, feedback is welcome. Look for some major feature announcements coming soon!
The Kontagent Team
Jeff Tseng December 28th, 2009 - No Comments »
Happy holidays from Kontagent!
For the last few weeks, we’ve been hard at work on this latest update, version 1.14. We’re happy to announce that we’ve released it just in time for New Years!! This version has some very significant updates such as a long list of UI tweaks, the added ability to support data export of graphs, upgrades to the test server and many bug fixes. Below you will find more details on each of these updates:
- UI Updates
- Re-written CSS style sheets that should speed up the load time and responsiveness of the dashboard
- Fixed issues with switching between application
- A brand new graphing package that should make it a lot easier to read the graphs
- Many data formatting fixes such as the addition of units and using appropriate number precision
- Numerous filter bug fixes
- Separation of time filter from the demographic filter which was previously causing conflicts with the preset filters
- Separation of traffic sources (ad tracking, partner tracking, viral sources) onto different pages
- Addition of revenue on the main dashboard page
- Test Server – One of the issues with the test server was that all applications for each account shared the same test server which meant that it was not possible for multiple teams to be testing different applications at the same time. In 1.14, we have updated the test server to facilitate one test server per application.
- Data Export -Data export has been one of the most commonly requested feature since we launched Kontagent. Now, all the graphs feature an export link in the top right corner of the window which allows you to export the data in each graph to CSV.
- Install Conversion Metric on SubType A/B Testing Tables – Release 1.14 features a new column in the subtype table call install conversion. This is the percentage of outbound messages that results in an install. This is a very important metric if you are tuning specifically for virality.
We hope you enjoy the new release. We hope you had a very merry Christmas and have a great New Year!
The Kontagent Team
Jeff Tseng November 25th, 2009 - No Comments »
Release 1.12 include a new feature called Integration Reports (available under Tools->Integration Reports) which is a tool that allows you to validate your live data to ensure your app is instrumented correctly. We have also added in feed channels to the Viral Optimizer as well as some fixes to the viral optimizer.
- Integration Reports – The test server was designed to allow you to test your your instrumentation in real-time, but it was not meant to validate large amounts of live data. We noticed that some developers were sending their live traffic to the test server to validate it, but the test server was designed to only hold up to 1,000 messages at a time. The Integration Reports addresses this need by allowing you to run a report that checks the last full hour’s worth of live data that was sent to our production servers. This enables you to see what we’re seeing as we’re processing the data. It’s highly recommended that you run the integration reports on your data to ensure that the data we are processing is as accurate as possible. The report can take a few minutes to run, so if you are sending us a lot of data, you may want to grab a coffee and come back to get the results.
- Feeds for the Viral Optimizer – If you now create a viral event using the Viral Optimizer, you will see a new option for feeds/streams/etc. in the event creation process. The metrics for the feed differ slightly from the direct messaging channels like invites and notifications in that the call to action and the messages are grouped together and must be tested together.
- Viral Optimizer Bug Fix – For some applications, you may have noticed that the 1 hour and 3 hour conversion rates for the message metrics in the Viral Optimizer management page were inaccurate. This has now been fixed, the 1 hour and 3 hour conversion buckets now display the correct values.
Happy Thanksgiving!
The Kontagent Team
Jeff Tseng November 6th, 2009 - No Comments »
Some fixes have been make to the new test server in addition to some cleanup of the REST API documentation. Deprecated parameters where removed from both the new test server and the REST API documentation. This does not affect current users. The fixes are as follow:
- removed ‘tu’ parameter from all messages except post messages
- removed ’s’ parameter in the post response message
- ‘r’ parameter is now optional in the psr message (it used to give an error if the ‘r’ parameter was missing)
- error checking added for the ‘i’ (has_app_installed) parameter
- removal of the ‘t’ template parameter (deprecated)
You may also notice that we’ve switched the Feedback tab on all the pages to the Zendesk version which integrated with our Zendesk support system.
Jeff Tseng October 30th, 2009 - No Comments »
As part of our effort to make integrating Kontagent easier, we’ve just released a revamped version of our test server which allows you to verify your instrumentation more fully. The biggest issue with our old test server system was that, while you could verify that you passed us data and what parameters we actually received, the test server never gave any feedback as to whether the message and the parameters were correct or incorrect.
We’re proud to announce today that we have launched a brand new test server that now includes error checking in addition to numerous other improvements. As part of the new interface, we’ve also made a list of all the different types of API calls that Kontagent supports and highlight whether or not that message type (invites sent, page views, posts, etc.) have been tested already and whether or not there are errors with these messages.
The new test server is available under Tools (main navigation bar) -> Test Server. Here’s a direct link to the test server as well.
We hope you find it helpful in the instrumentation process. As always, feedback is welcome.
Jeff Tseng October 26th, 2009 - No Comments »
It”s finally here! Feed A/B testing! You guys have been asking for it forever and we finally got around to getting the A/B testing table working for all the post type channels such as the Stream, Feed Story, Publish User Action. This means that you’ll be able to use subtypes to categorize your feeds and A/B testing them accordingly.
We also got a request to have a graph of installs vs. time for the post channels and as well so we’ve that as well.
On a separate note, we recently added a new page to the developer site that outlines the Kontagent A/B testing framework. Check it out if you haven’t started using our A/B testing framework yet.
Stay tuned, we have another update coming very very soon!
Jeff Tseng October 26th, 2009 - No Comments »
As you’ve probably noticed, we’ve overhauled the MyApps page. One of the issues we were always having was that we had no way of updating you (the developers) on the many new features and fixes that we’ve been making to Kontagent. With this new release, we finally have somewhere to post updates about the product which we will now do on a consistent basis.
In this release, we made the following updates to Kontagent:
- Added the Getting Started and Updates section on the MyApps Page.
- Moved the comparison graphs of all the applications into the Compare Tab.
- Added an Updates page which will contain a full list of all the developer updates.
- User Management Fix – Some developers were experiencing an issue where only the original account creator could create applications. This has been fixed. Other admins and users with app management access can now add applications.
- The percentage change in the Daily Uniques and Installs columns on the MyApps page is now the change compared to the same day the previous week (7 days ago).
Look for plenty more updates coming soon!
Jeff Tseng October 26th, 2009 - No Comments »
Release 1.7 includes two fairly major feature additions: Monthly and Weekly Retention Analysis and a new User Management System.
Retention Analysis – User retention metrics are critical to understanding how engaged your users are and how to improve your application’s user lifetime. This new features provides month over month percentage returning users grouped by monthly installs cohorts retention metrics and week over week returning users grouped by weekly install cohorts as well. Retention Analysis is only available in the Premium Analytics package.
User Management – The new user management feature allows you to grant extremely granular access to each of the users in the system. You can set up the system to show only certain tabs to certain user, or grant users full administrative access. To access the new User Management system, to go “Account“->”Manage Users“. Please note that you can also access this tab if you have account management privileges.
