Continuing our market research for the JobFinder App (as described in my last blog) I decided it would be good to examine the analysis that has been generated via Flurry. Looking at the past usage of the app should (hopefully) give us an insight into what currently works well within the app, and thus should be continued, and what doesn’t. The theory sounds all well and good, however in practice it is not that simple.

I can see from the statistics that there were 8,096 sessions within my chosen time frame (a session being from when the app was opened to when it was closed). I can also see that those sessions were actioned by 1801 users in that period. Of which 886 were retained users. However when I try to analyse what those sessions involved it is a little more difficult.

Flurry gives us some information on performed searches so we can see the type of vacancy sought and the location of users who searched. However the data isn’t pre-analysed except to show that there is very little consistency in what or where users search for jobs.

Screen Shot 2012-05-10 at 16.12.40

In order to obtain some more meaningful data it is necessary to download the event logs (a page at a time) and perform some offline analysis. I did this on the last 50 job searches, I know that isn’t really a fair representation but it still gave me some trends.

Country UK
Vacancy Location Area
Named Place
Post/Zip Code
Not Specified
Vacancy Type Industry
Not Specified

I concluded the following:-
Location name was the most popular way to search. This could have been manually typed or could have been generated through the apps ‘locate me’ function, there is no way of determining that currently.
. Some users want the ability to just say show me all jobs in my area
. Some users want to be able to specify a company rather than a job name
. Some of the named roles would require a more specialist agency search, e.g. babysitter
. Whilst others give very vague terms such as Manager or IT

There were also a lot of repetitive searches suggesting that saved searches and the ability to just refresh the results on a new day would be a useful feature.

The other area within Flurry to look at is audience demographics. We don’t ask for information regarding a users age or gender within the app but Flurry does provide some estimates based on an aggravate of all of the data that they collect. The estimates didn’t really show any surprises as we expected the age range to match that of the working population. However there was an increase in users within the 18-24 age group, potentially due to the number of school and college leavers looking for work. There was also a significant decrease in user numbers in the over 55 age range. However this may change as more tech savvy users enter this age bracket, and according to a recent report over half of recent UK smartphone purchases were made by the over 50’s. With the planned increases in the retirement age and predicted pension problems this is certainly a user group that can’t be ignored.

For our app the geography statistics are far more accurate demonstrating that the majority of the app users are in Europe (51.2%) vs North America (48.4%). However what is surprising is that there are some figures (albeit small ones) for other territories such as Asia and the Middle East. This is despite the fact that we haven’t localised the app into other languages, and more importantly it is only available in the UK and USA app stores. To expand the app into other territories or even into other countries within Europe, we could not simply localise the app into the native languages of those countries. We would need to determine the typical routes users in those countries take to find and apply for jobs. We would need to expand the list of sites we source jobs from as well as determine if there are any nuances or restrictions those countries apply when jobseekers are searching for jobs. We may well expand our reach into other English speaking territories but localisation is more likely to be something we look at in the future. For the next version of the iPhone app (and the first version for Android, Blackberry and maybe other platforms) it is not something we want to concentrate on.

The more interesting demographic for us is the category listing. Flurry uses its amassed data across a multitude of apps to give an indication of which category on the app store most users spend their time. Not only does it give a breakdown of the benchmark of all user sessions against category, it also gives a breakdown of our app users sessions within each category. Currently JobFinder is in the business category within the app store with a secondary category of Lifestyle. However when you look at which category the apps come from that users spend most of their time in, it is the Lifestyle category (after Games and Social Networking of course!). The business category is 22nd on the list. This suggests we should switch the categories around. However you have to look at the other side of this argument. There are currently over 17,500 apps in the Lifestyle section on the UK iTunes store and only approximately 5,600 apps in the business category. So there is a lot less competition in the business category and thus far more chance of getting your app noticed. Does the chance of a top hit in the business category outweigh the opportunities of getting seen in the lifestyle category (as it is far more visited). Or does this now hold little relevance? Do most users look through each category to find out whats new and noteworthy or do they simply search for names and related keywords on their device and download from there? To try to find out the answer to this I have decided to create a LinkedIn poll and see what responses I get. The survey has quite a long time to run but I will let you know any results in a subsequent blog.