We at Adjust have some really strong opinions about what an attribution solution should be. As such, we’ve realized that any attribution platform needs to include some baseline essentials that don’t overlook the core needs of attribution. But what are they, and how important are they in reality? Let’s take you through what we think are must-haves to any and all attribution platforms.
1. An open-source approach
An open-source approach is the only way to work, whether now or in the future. With an increasing amount of businesses, institutions and even countries mandating open-source as the means to create and publish, there’s no reason to begin your project as proprietary software, nor to justify continuing to do so. While some attribution providers may claim to be transparent, this cannot be the case without open-source at the core of the product.
2. Indefinite look-back on your segments
It’s quite simple really, in spite of the imposing title: a good attribution solution won’t delete certain kinds of data, no matter how long it’s been around for (though they will delete others, more on that in a second). Having a year, or two years’ worth of data around segments or cohorts is incredibly valuable, and makes planning for a campaign that much better for having it. You shouldn’t be hindered in viewing historical data by any shortcoming of the platform, nor face a nasty surprise from being unable to find your records, only to discover that they’re gone forever.
In our case, we don’t store raw data on specific user’s actions indefinitely, as this would go against European privacy compliance and also certain high profile partnership programs. However, we do carry on aggregation forever, such as for cohort analysis, or for attribution. And we don’t forget users after a fixed amount of time - if we did, it would spoil your data.
3. Unlimited API requests
Similarly, using APIs shouldn’t be limited. Any attribution platform should allow you to make as many requests as you like, regardless of the size of data you want to transfer or download, or frequency of API calls that you make.
Rate limiting is very common in the industry, and there are providers that ‘throttle’ the number of queries you can make. This is like Google only allowing you to search twice a day, or download ten results per request. If you’re looking at solutions, ensure you check-up each providers’ API policy, because you should be able to take the data you want, when you want it, without limit.
4. Privacy compliance
Privacy is a big deal. It’s a constant and prolific concern for both businesses and the public, so ensuring your attribution solution is compliant with privacy laws is hugely important in this day and age.
One such assurance of privacy compliance is to be certified by either the ePrivacy group in the EU or the Privacy Shield group in the US. You can go further too, in Europe at least, by being compliant with EU law (which you can find out more about here). Such actions would be imperative for any digital service in showing how serious they are about privacy.
An attribution platform that avoids third parties, and cloud storage, should strengthen your confidence in them. Meeting these criteria can be tough, but if you are wary of privacy (and the laws that certain attribution partners could fall afoul of) then it’s important to research the policies of each respective company when making a decision.
5. Good relations with major players
Google, Facebook and Twitter make up a huge majority of the digital ad marketplace, and in order to access their data your attribution provider needs to be part of their measurement programs. By being a part of such a program, your attribution platform will have access to data from some of the best performing advertisers in the world, right in the dashboard. On the flipside, by being on the outside, it’s impossible to analyze campaigns from many of these networks. This means the product could have demonstrated much lower standards than acceptable.
If your attribution provider isn’t a part of a program then you must consider why this might be, what might it mean about their software, and what roadblocks this could present.
6. In-the-moment fraud prevention
Fraud prevention is only prevention if it stops fraud in its tracks. Mobile ad fraud is a persistent and hard to beat problem in the industry. All providers ostensibly offer some kind of fraud prevention, but the trouble is that many of them don’t actually prevent fraud, just inform you that it’s taken place after the fact. By letting fraud happen, not only will you lose revenue, but problems between you and networks could occur. Furthermore, your datasets will become skewed, and made practically redundant, by the presence of fraudulent activity.
If fraud is a top priority, you need to ensure that your chosen solution is actively stopping criminal activity from affecting your bottom line.
7. Organic events in a cohort
While much of what you’re tracking will be to do with paid campaigns, it can still be very useful to know where your other sources of traffic and transactions come from. As opposed to an uncertain figure representing all activity outside of paid sources, having organic traffic listed within your cohorts clarifies much more about what’s going on outside of your campaigns, while allowing you to identify potential new niches, and giving you the ability to grab some quick wins.
8. Comprehensive account history
A detailed account history that shows you who created tracker links/campaigns, and when they did so, can help improve internal accountability, and also help when it comes to optimizing your future campaigns. Some dashboards can track your installs from search, but some don’t, while others may ask you to copy and paste your data into a document or spreadsheet to save this information. The best attribution solutions should keep all of this information for you on the dash.
9. Real-time data
It’s incredibly useful to have all your data up-to-the-minute, as opposed to syncing every so often. If your cohort data is calculated only once a day, you’re potentially missing out on a range of crucial datapoints, which could make last-minute analysis or quick presentations that much more difficult to make.
Does your attribution solution cover off all of these points? If not, you may want to consider trying adjust for yourself, just so you can reap the benefits of a complete attribution platform.