Building a Best-In-Class Stack for Truly Innovative Messaging Campaigns
While choosing the right messaging platform is critical, how you integrate it into your business is a bigger indicator of your ability to drive innovation and impact. Last week at MAU I moderated a panel that discussed the use of machine learning in messaging. Given the amount of interest in this topic, I thought I’d take the time to delve deeper into the topic by sharing some integrations that will empower your team to start running some truly innovative messaging campaigns.
Bringing your messaging platform online
First things first: if your goal is to prioritize broader CRM efforts over mobile messaging, then you’ll want to skip ahead and focus on integrating your data warehouse. However, if your goal is to build a robust mobile messaging campaign, one of the first things you’ll want to do is add the SDK to your website and mobile app(s). Not only will this enable you to send in-app, web and push notifications, it will also allow you to create cohorts based on front-end user behavior and tie downstream engagement and performance to specific campaigns.
Unless your engineers love SDK integrations, I’d highly suggest utilizing a Customer Data Platform, or CDP. Once the CDP is implemented, it functions as the single SDK on the front-end, using its own backend to distribute data to other platforms that need it. For example, in this instance, we’d directly integrate a CDP into our website and apps, then “activate” our messaging tool through that. Not only does this save your engineering team a lot of time and energy, but it also reduces your marketing team’s dependence, giving them more flexibility to implement the latest technology.
Connecting to your data warehouse
The right setup can take your marketing team’s ability to be data-driven to a whole new level, and it starts with your data warehouse. Connecting your messaging platform with your data warehouse will enable you to receive more precise transactional data in instances where you might lack visibility on the front-end, like actions that occur across channels—even those that take place offline. For example, let’s say you’re hoping to further engage your brick-and-mortar customers by sending them a confirmation email after they’ve made a purchase at a physical store. You wouldn’t want to trust front-end analytics with a task such as this, but the backend integration will get all this data in for a nice, timely message once the sale has gone through.
The value of third-party analytics tools
While your data warehouse should always be the ultimate source of truth, it probably isn’t intuitive enough for many members of your product or marketing teams to utilize—this is where third-party analytics tools come into play. By integrating your messaging platform with a product analytics tool, you’ll empower your marketing team to conduct the impact analysis themselves without having to rely on a data scientist.
If your messaging platform provides built-in analytics, you should still connect to a third-party analytics tool. The fact is, many of today’s most powerful messaging platforms simply aren’t built to do double-duty as in-depth analytics tools. This is why most messaging platforms open their APIs to more comprehensive product analytics tools and data warehouses. Braze and Amplitude are a great example of this—the two platforms have a deep partnership that enables marketers to create a full loop of data from one system to another. You can use this integration to take cohorts from Amplitude and import them directly into Braze for retargeting while continuing to analyze in-depth campaign performance in Amplitude, taking your marketing team’s ability to be data-driven to a whole new level.
Using a Customer Data Platform for truly innovative campaigns
I’ve already touched on how a Customer Data Platform can reduce your team’s reliance on engineering for SDK installations, but that’s just the tip of the iceberg. In addition to the aforementioned technical efficiencies, a CDP can connect to virtually any system you might have—frontend, backend, ad networks, marketing automation, business intelligence, and so on.
Once properly set up, your CDP will be able to pipe data with minimal engineering time, enabling you to run interesting campaigns that will delight your customers at every turn. For example, you can utilize a CDP to sync your crash reporting tool and messaging tool to send potential customers an apology and discount code whenever they experience an app crash, turning a frustrating situation into a positive customer experience—and more than likely, a sale.
A best-in-class stack makes your marketing 🔥
The real beauty of this setup is that you can create a user cohort directly within the CDP and keep that user cohort synced between many systems. For example, let’s say you’re working between four platforms: a messaging tool, a CMS, an analytics tool, and Facebook. You can use the analytics tool to create a segment within your CDP (ex. “all users who have viewed 3+ products, but haven’t been active on any channel for 15 days”), and sync this cohort in real time across all platforms for a cohesive, uniform campaign that includes email, push notifications and retargeting on Facebook. You can even use your CMS to serve up tailored content for when those customers finally return to your brand. Really powerful stuff!
Two of the more popular CDPs are mParticle and Segment, both of which we use quite often at Prolific. If you’re not already familiar, I’d recommend exploring mParticle’s AudienceSync tool, which enables you to define audience segments based on any combination of attributes, layer in third-party data, and seamlessly sync these segments in real time across virtually any platform. I’d also recommend that you check out Segment’s new Personas tool, currently in Beta. Personas unify customer history from multiple channels and devices, building a robust profile that enables you to deliver personalized marketing campaigns and in-app experiences.
Taking things to the next level with machine learning
There are some really exciting machine learning applications that are easier to implement than you might think. If you’ve already connected your messaging platform to your data warehouse and have a good partnership with your data science team, you’re halfway there! An easy place to start is by leveraging product affinities in your re-engagement campaigns.
To get started, ask your data science team run a machine learning (ML) algorithm to determine what a customer is most likely to buy based on a set of user behaviors and/or past purchases. Once the ML algorithms mature a bit, turn them into cohorts that can be fed into your messaging platform directly or via your CDP. With this in place, you can now use those cohorts in your messaging tool and target those users with the exact product, or product category, that your ML algorithm predicts they’d like to buy. Easy, yet highly effective!