.svg){kind=small}
Thanks For Subscribing To Our Newsletter
Something went wrong. Please try again.
Book a meeting with one of our marketing consultants and see how we can help you improve your ad spend and increase your LTV.
Thanks for scheduling a meeting with our team, we're excited to show you what we bring to the table.
Read a case studyBy now, most marketers have come to the realization that a new user-acquisition era is enforced on us all, tracking has changed forever and optimization based on the short-term conversion window is on its deathbed.
And while it causes a lot of confusion, it also holds an opportunity: In many cases, the short-term conversion window has almost no bearing on a user’s true LTV, regardless of the upcoming global release of iOS 14.5 that forces the entire industry to rethink campaign optimization. We see these changes as an opportunity to evolve and improve the way marketers go about the complex game of user acquisition (UA) campaign optimization.
For many growth marketers wielding huge UA budgets, especially those hailing from SaaS, B2C, D2C, subscription-based, Fintech, Gaming industries, and similar, the stakes are high. Most already know that they need a new, robust framework to help them deal with the upcoming challenges. They’re well aware that the new name of the game is predictive analytics and optimization technologies that help brands target the most sought-after audience — high LTV users or users that are simply best for the business.
Leaders of the growth marketing pack are on the lookout for related technologies & solutions that will help them send relevant user-level optimization signals to the ad networks and by that not only bypass the current iOS14 issues but also reach new standards of optimization.
But, finding the ultimate solution — for UA or in general — is no easy task in such a technical environment.
Well, we, therefore, used our entire Voyantis team and some growth marketing gurus, to put together a list of 9 questions to use when implementing a UA predictive solution, in order to understand which is the right one that fits the team, budget, scale, internal resources and more.
Ready? Here we go!
Predictive UA solutions are basically fed by data, but that doesn’t mean you need to be a data monster. The AI technology that drives predictions needs to be fed historical data in order to produce accurate results. Now, if the media budget is low, this learning phase can significantly reduce the ROI of such solutions. However, companies that spend around 100 thousand dollars a month on UA will see amazing results in a very short time, with these bleeding-edge solutions.
Think carefully before implementing a solution that simply leans on historical averages. Historical averages of short-term events and static data like Country/Ad network/Creative type, etc., or even regular cohorts, won’t generate accurate enough predictions that can deal well with today’s UA challenges. Such historical averages and macro-level data provide limited-efficiency deployment of marketing capital. When focusing on a country or city cohort, one may miss out on diversity within the sub-segments and on critical individual-level indicators that aren’t tracked within such limited cohort segmentation.
Also, that’s what everyone is doing as low-hanging fruit — and this is not a competitive advantage.
We’ve already discussed the problem with short-term focus; it fails miserably as an indicator of LTV — just think about a subscription or other recurring purchases, or long funnels, or free trials.
New-age predictive UA solutions overcome these limitations elegantly, by leveraging user-level predictions for a holistic view. Individual users' data enables highly accurate predictions per each user, which further allows using the predictions in any version of a slice-and-dice. Imagine defining a cohort as anything between a single user, i.e. a “group of one”, and a “regular” campaign-level daily cohort without the limitations of a pre-defined cohort based on the average of a large group.
To complete the picture, ad networks are ready to receive such signals from the buyer's data systems, in order to optimize UA. Data ownership shifts (or is rather pushed by ATT, iOS 14.5, and the Cookieless trend) towards the buyers. Facebook’s new Conversion API, for instance, allows predictive analytics solutions to share server-to-server, user-level predictions to enhance optimization capabilities in the campaigns that are run on its platform. Google has developed a (sort of) similar solution.
The problem with many of the predictive UA solutions out there is that they require lots of the customers’ engineering time and effort to get them up and running. And that’s just the setup; in many cases, these solutions demand significant attention from engineering to make even the smallest changes and to actually use them. This issue can become a deal-breaker for even the most powerful predictive solution, as it can often take months to organize the resources required to get moving. In the majority of cases — with only giant companies being the exception to the rule — companies don’t have spare engineering capacity, and engineering becomes a bottleneck that prevents marketing from achieving their goals.
Forward-thinking brands are rising to this challenge. Vendors are offering no-code solutions that are easily set up and maintained without requiring any resources from engineering. Growth professionals are looking for no-code alternatives that provide them with tech prowess without having to involve R&D, as they want their goals to be met in a reasonable time.
Not every predictive UA solution is right for every business model. This is because the specific type of business model used by a business affects the way that LTV is measured. Some business models focus on subscriptions, which by definition take more time to measure LTV. Luxury or premium products (or services) vs. instant purchases also affect the measuring mechanism, as the buying journey is usually much longer when it comes to luxury purchases.
The same holds true for the rest of your marketing solutions. Not every predictive UA solution necessarily works well with the specific 3rd party data sources and marketing automation platforms that you’re using. Each predictive UA solution is built with a certain set of use cases in mind, and is built for easy integration with certain platforms.
Bottom line is to double-check that the predictive UA solution you’re considering can support your business model and integrate with the rest of your martech stack.
Not every predictive UA solution provides the same output. While some solutions provide general insights, others provide actionable optimization recommendations that can immediately be used to boost performance. Platforms also provide different levels of granularity in regards to the insights, recommendations, and actions that they offer.
While some offer just the predictions (models), others offer insights and actions (export and use) via API. In the next tier, automated actions are offered. Our advice is to carefully check which tier is offered by the specific solution you’re considering.
A predictive UA solution is only as good as the data you feed it. That’s why you should always check that the solution you’re considering goes beyond 1st party data (your owned data) and can be enriched with 3rd party data. These external data sources add critical data layers that significantly enhance accuracy levels. Some examples of 3rd party data types can include census, location, weather, and events/holidays data. Of course, data privacy and compliance issues (such as GDPR) should be considered.
At the end of the day, a lot comes down to budgets. Before you start onboarding your new predictive UA solution, make sure you understand exactly how long it will take before you see actionable results that can be used to generate better ROAS, optimize business results and prove the ROI of the solution.
The best predictive UA solutions were built for the needs of growth marketers (rather than BI or RND teams). They talk the marketing talk and walk the marketing walk by implementing friendly features and complying with marketing performance terminology like CAC, ROAS, ROI both within the UX writing and as metrics for success. If you’re looking to optimize your growth marketing, stay away from predictive UA solutions that were built for data scientists or for technical teams, and measure success with prediction metrics like precision and recall.
Last but definitely not least, how is the elephant in the room called iOS 14.5, the one we mentioned above, addressed? It’s a real game-changer that is upending the way many performance marketing platforms operate. Ask how the solution addresses the change and if it’s limited to the SKAD Network? In a nutshell, the monumental changes brought on by iOS 14.5 dictate that brands learn how to navigate the delicate balance between user quality and quantity. They’ll need to learn how to run and optimize campaigns that are triggered by user value, instead of user actions.
Check out this article for the full rundown of our take on the implications of iOS 14.5.
Whether you like it or not, change is a constant in our lives. This time, it’s growth marketers are in the cross-hairs with the decline of the short-term conversion window coupled with the launch of iOS 14.5. New, predictive UA platforms that leverage predictive analytics and optimization techniques offer growth marketers an innovative framework for this brave new world.
However, the shrewd marketer should always look before they leap, and ask the right questions to make sure that the solution they’re considering meets their needs.
I hope that the list I’ve compiled above helps you on your journey to find the perfect predictive UA platform for your specific needs.
