Nurture, Optimize, and Deliver High-performing Campaigns with Algorithmic Modeling
Adobe Audience Manager empowers you as a marketer with unprecedented access to valuable information about the prospects interacting with your brand. This information is assimilated by segmenting the unified first-party online and offline prospect data that you own, as well as the second- and third-party subscription data from partners and data providers available in Audience Marketplace.
This information includes demographic, psychographic, as well as behavioral aspects of the prospects that you can use to create actionable audience segments. But when coupled with algorithmic modeling (powered by Adobe Sensei, Adobe’s artificial intelligence and machine-learning capability), these segments of audiences present a tremendous opportunity to create data-driven target customer segments. These segments empower you as a marketer to identify, refine, and nurture your marketing campaigns at every step of the campaign journey.
In this blog post, I will discuss some opportunities where the look-alike algorithmic modeling feature in Audience Manager can help you nurture and optimize your campaigns for stunning performance.
Use Predictive Analysis to navigate your path to campaign success
Reaching the right audience at the right time with the right message are key drivers for running effective campaigns. Adobe’s algorithmic modeling features and capabilities allow you to use a data-driven approach to generate effective underlying audiences that you can nurture at every stage of the campaign to run a highly optimized and successful campaign.
1. Find new prospects/audience that look like your existing customers
Finding new and relevant prospects to expand the customer base is expensive, resource intensive, and time-consuming. Also, finding them in time before your competition does is a huge challenge. Using the algorithmic model capability, powered by Adobe’s proprietary TraitWeight algorithm, you now can find new audiences from subscribed Audience Marketplace data providers and partners’ universes that look like your existing customers.
All you need to do is provide a trait or segment of your existing customers as input to set the baseline for comparison in the model and it scores the audience using a machine-learning algorithm (TraitWeight Algorithm) on Reach vs. Accuracy scale.
The model output then lets you choose a subset of a large pool of discovered audiences. You can then start to target these audiences for your campaign via marketing channels of your choice.
2. Nudge the prospects to advance in their customer journey to make them sales ready
Once prospects enter the marketing funnel, depending on the stage they are in, you can engage and drive them to the next stages of the journey (e.g., depending on the size and budget, you can orchestrate separate algorithmic models).
For example, for prospects at the Awareness stage, you can create an algorithmic (look-alike) model in Audience Manager to find audiences that look like the previously converted audiences or audiences already at the Interest stage. You can do that by making a selection at Accuracy vs. Reach score scale and then retarget them with appropriate marketing messages to help push them to the Interest stage. By doing this, you are moving away from “spray and pray” to a smarter, measured, and narrower data-driven targeting, saving yourself a chunk of money on the marketing effort coupled with larger probability to engage.
For the prospects that are now at the Interest stage, marketers can create another algorithmic (look-alike) model to identify the most valuable audiences that look like their existing customers to retarget and help them reach the Consideration stage.
For the rest of the marketing funnel stages, such as Intent and Evaluation, where the volumes are manageable, you can either segment on other data points to identify optimal audiences to retarget or continue using additional look-alike models depending on the size of the target volume.
The idea is to nurture the most qualified audience likely to move to upcoming stages of their customer journey with relevant offers and experiences they would respond to. This drives the cost per acquisition down and improves conversion rate due to smarter, AI-driven, targeted audience selection.
The look-alike model, powered by Adobe Sensei, is accessible directly to marketers as an Adobe Audience Manager feature. With this feature, Adobe is bringing the artificial intelligence capabilities into the hands of a marketer. The interface hides the scary complexity to a very large extent by providing the only levers necessary to supply inputs and understand audience outputs, making it extremely easy to set up and use. “In-built automation” is another mention-worthy capability of the model that allows it to continually learn as and when new data hits the system without the need for any manual intervention. This helps you achieve marketing automation with a few clicks.
There are several other use cases for which marketers can take advantage of this feature (e.g., better content distribution and empowering cross-sell or upsell). For publishers, the premium use case is to find look-alikes to their premium audiences and monetize these audiences. Both marketers and publishers can benefit from look-alike modeling to understand more about their high-value audiences by studying their most influential traits exposed by these models.
If you have not flexed the muscle of machine learning-based audience selection, now is the time to start using it to benefit your campaigns and organizational goals.