DM: So Jo, there’s a number of confusion around tracking of ads in-app and at the mobile web and there were about a changes of late, what’s the present state of play, and maybe you’ll want to take it one platform at a time, so let’s start with iOS.

JW: Sure, so greater than twelve months ago, Apple said it not desired to see UDID, a static identifier that’s unique to the device, getting used for tracking clicks in apps and app downloads. But they did this without popping out with an alternate, and this created confusion and disruption. So things then fragmented as people scrambled to locate a replacement, so we had Open UDID, the MAC address and HTML5 first-party cookie tracking. 

Then with the discharge of iOS 6 in October, Apple released the Identifier for Advertising (referred to now because the Advertising Identifier or IDFA), as their sanctioned replacement for UDID. It’s reminiscent of UDID in lots of ways, but crucially, the user can wipe the IDFA and opt out of it.

DM: So what has the uptake been like for IDFA

JW: We didn’t see immediate adoption. It came into some SDKs (Software Development Kits) however it was not as though everyone immediately converted to it. But we’ve seen some acceleration because the 1 May deadline for developers to forestall using UDID was announced on 21 March. The announcement of the hard deadline had the effect we might have anticipated, in provoking a shift towards the advertising identifier. As of mid-April, 70 per cent of the inventory that may be tracked is being tracked by IDFA. Also, the complete volume of inventory being tracked is up. And we see an analogous choice of ad networks supporting IDFA that in the past supported UDID. 

DM: And what of those interim solutions now that the 1 May have been and gone, what’s Apple’s attitude towards them

JW: They’ven’t made any statements on Open UDID. They have also not said anything formally on HTML5 first-party cookies, but I actually have heard they don’t like them, and that i hear an analogous in regards to the use of the MAC address.

DM: And from an advertiser perspective, are there any issues with using IDFA in place of UDID

JW: Not likely, there’s been no drop in performance, and no indication that the facility to the wipe identifier is having an impact on tracking. It’s been relatively easy to implement from an infrastructure standpoint.

DM: OK so we all know where we stand with iOS, what about Android

JW: With Android, it’s pretty plain sailing. They use a system of referrer tracking, using a URL to which different parameters will also be appended, like web-based tracking. It’s as much as the advertiser what they capture, reminiscent of the time of click or the IP address. It gives really good attribution, and there are not any issues or concerns around a non-public identifier.

DM: OK, and what about clicks on ads at the mobile web

JW: We use digital fingerprinting. This can be a URL that captures a lot of attributes concerning the devices, the time of the clicking, the IP address, and we use statistical techniques to compare at the back end and make a determination of the conversion rate.

DM: And the way accurate is all this stuff

JW: It’s all 100 per cent accurate except for the digital fingerprinting stuff at the mobile web. We’re seeking to get deeper than simply tracking the clicking and the install; we wish to determine the ROI and marketing spend, so we have to know what happens after the install so as to exactly make that match. Digital fingerprinting is perfectly adequate for clicks and installs, but less helpful for the deeper analytics.

DM: And what about other platforms, BlackBerry, Windows Phone, Facebook, Twitter

JW: As far as we’re aware, there is not any methodology to trace for Windows Phone or BlackBerry but they can use something like Digital Fingerprinting, and that i think here is what Amazon will do. Facebook has its own unique identifier that permits for accurate tracking. Twitter, I haven’t heard of anything like that but they will have plans.

DM: And simply to talk stats for a moment, are you seeing any trends on the subject of a shift from mobile web to apps or vice-versa

JW: No real shift occurring. Around 30-40 per cent of the inventory we’re tracking is mobile web, but i feel the mobile web is under-utilised by app marketers because they generally tend to default to in-app.

The variety of apps being released remains to be on an upward trend. There have been 20bn app downloads by the tip of 2012, and that’s expected to achieve 40-50bn by the top of 2013. We’re expecting the variety of Android app downloads to overhaul iOS in this year. And when it comes to the collection of apps within the app stores, as of last month, there have been 830,000 for the iPhone and iPad touch, 300,000 for the iPad, 680,000 Android, and by means of comparison, 130,000 for Windows phones.

DM: And is there the rest within the offing that developers should be familiar with.

JW: Nothing stands proud. We’re delighted to have to this point. We’ve expended a variety of energy and put various effort into managing the UDID transition. We now have done three or four SDKs in the course of the period to offer clients maximum flexibility, so we’re pleased that we will be able to now put that behind us and concentrate on a better thing, and hopefully, we can see more uniformity and not more fragmentation, and this may open up possibilities to do better targeting and retargeting as people feel confident that privacy controls are in place, so there isn’t a cause of any reluctance to develop more targeting options, and to retarget individuals who have already got the app installed to re-engage with it.

One trend we’re seeing is developers making more of an effort to realise the various cohorts of people who download your app, and what action they take after downloading – registration, repeat opens, but in addition setting out to understanding engagement through the years, purchase behaviour, and rates of social referral. We’re also seeing a brand new group of individuals doing user acquisition for apps who come from digital, in preference to mobile backgrounds. They may be observing the distance, comparing it to desktop, and looking out at what can they do similar on mobile, and the budgets listed below are often associated with more established brands, similar to retailers etc., but a few of bid speccing behaviour would possibly not look rational to an app marketer.

Finally, because the tracking issues are ironed out, here is opening up more targeting options, and we predict this could benefit European advertisers, on the subject of isolating particular countries and cities they’re more interested by targeting. Europe is a really fragmented marketplace, in comparison to the U.S., so targeting becomes more important.

Jo Wightman is director of client development at Fiksu