Mobile attribution: the challenges and how to get it right

Accurate mobile measurement is essential to assessing the ROI of your mobile advertising campaigns, but it is very difficult to do properly and across the entire mobile ecosystem – both in in-app and mobile web environments. Cross-environment measurement is complicated because each silo lacks common tracking and has varying degrees of functionality.

Attribution on desktop is simple and proven, as it relies on cookies. However, that method doesn’t work in mobile. So, how do you get around it? Without a universal method to track campaigns across mobile devices, the efficacy of mobile campaigns can be underreported and devalued.

There is a lot of noise about mobile attribution, but do we really know what it all means? Do we know how to look out for the black holes and what it will mean to your reporting?

The mobile issue

The first issue is that the industry is still using old-fashioned desktop metrics to define success. Outside of brand campaigns, click-through rate is still the most common measurement of success in mobile advertising. Until very recently, only click-through attribution has been possible, which in a world where everyone is talking about big data is very underwhelming.

The measurement of success on your mobile advertising campaigns should be attribution – both view-through and click-through.

Utilising attribution properly should:

  • Give you total visibility into post-impression in-app or mobile web activities that are most critical to your campaign. Examples include product purchase, request for more information, app download and activity, etc.
  • Tell the full story of your campaign success by tracking both view-through and click-through activity. Track attribution for your entire campaign, including ads that run on different devices, publishers, and ad networks – across both in-app and mobile web environments.
  • Optimise campaign performance by attributing post-impression activity for different publisher properties, creative, and ad sizes.

What works on desktop doesn’t work on mobile

Attribution is simple and commonplace in desktop environments. The industry recognises that desktop inventory sources use cookies as its main signal. However, cookies are not persistent in mobile, so a combination of other methods is needed to identify the mobile device.

Deterministic and probabilistic identifiers are the two main methods used for attribution in mobile. Many vendors provide either one or the other, and by only using one method, you will not be able to fully identify a device across both in-app and mobile web inventory sources. Ultimately, you will have incomplete data.

What do they mean and how do they work?

Deterministic: A deterministic identifier, such as a device ID, is a guarantee that the user is unique. However, this is only possible in in-app inventory and only when made available by the inventory source (publisher, network, exchange). Not all inventory sources share a user ID or a device ID. Running a campaign on mobile web inventory means you cannot collect the device ID.

Some companies, such as Google and Facebook, use your user data when you are logged into their apps. If you are not logged into their apps and they are not running in the background, they cannot obtain a deterministic signal.

Probabilistic: A probabilistic identifier is a piece of information that, on its own, cannot uniquely identify a single device. Examples include user agent, browser version, time zone, etc. While probabilistic identifiers are less accurate than deterministic identifiers, they do work across in-app and mobile web inventory sources.

What to watch out for:

  • ‘Walled gardens’ can offer attribution, but only if a user is logged into their system. Walled gardens do not share their signal data.
  • A device ID is the most accurate signal that identifies a unique device, but is only accessible in in-app inventory and only if the inventory source shares it in a readable format.
  • Many providers offering attribution need the user to complete the conversion in a ‘single session’. If the user closes their webpage or doesn’t complete the transaction in that session, the impression will not be attributed.
  • Most providers can only track and offer attribution either in in-app or mobile web inventory, but not in both.

When you get it right:

  • Collecting view-through and click-through attribution will add another dimension to your data. The holistic data empowers marketers to measure and optimise ROI on every mobile impression across every app and mobile website.
  • Including comprehensive conversion details, such as product type, cost, and order ID, will allow marketers to understand user behavior in mobile environments.
  • When a probabilistic identifier is combined with a deterministic identifier, the two methodologies together give a high degree of certainty while providing the most level of coverage and scale.

What you need to look for:

  • Be sure to collect attribution across in-app and mobile web environments. Ask your provider if they can track in both environments before the campaign goes live to ensure you’re able to measure ROI across all inventory sources.
  • Confirm that you can collect both view-through and click-through attribution in in-app and mobile web environments, allowing you to examine the full ROI of your mobile campaign.
  • Check that the provider can pass unique variables (such as order ID, etc.) through the conversion tag so that you can verify the purchases and conversions.
  • Use a third party vendor for validation. You would always use a third party vendor for your desktop buy, so be sure to do the same in mobile.

Richy Glassberg, CEO, Medialets WW LLC

Tel: 212-300-5670

Email: connect@medialets.com james.oconnor@medialets.com

Web: www.medialets.com

Twitter:@Medialets

Get The Drum Newsletter

Build your marketing knowledge by choosing from daily news bulletins or a weekly special.