Advanced Retention Strategies for Apps
Savvy marketers know the playbook for a winback and a welcome series, but how do you consider that against a multi-channel, mobile first environment? Dive in for more advanced strategies to level up your retention campaigns.
Steven Aldrich
A portion of this content was originally presented on June 22, 2023, at the App Promotion Summit.
Overview
For app-only businesses, this download-delete-redownload challenge is greatly exaggerated and can be detrimental to continuing user growth. For the most part, the strategies discussed in this article can apply to app-only and hybrid businesses. Depending on the audience of your app, such as gated access until purchase, some of the timing will need to be adjusted to coincide with the user’s overall experience.
Before we jump into strategies, let’s quickly align on the definition of the audience lifecycle:
Leads / Prospects
- New: Within the first 7 days of entering your app and have not converted
- Non-X: More than 7 days, have not completed your defined conversion action (purchase, subscribe, listen, etc.)
- Lapsed: More than ~30 days, have not completed a converting action with your app or brand
- Sunset: More than ~90 days, are no longer relevant to your brand and have an extremely low likelihood for revenue
Converters
- New: Converted for the first time within the last 7 days. You define conversion as a user who completes registration on a freemium model, or someone who pays for a premium offering. Regardless of which you use, driving engagement in their initial period will support longer-term retention.
- Post: Converted more than 7 days ago. Duration dependent on when the user is eligible for additional conversion activities (typically 30 days for most brands)
- Re-Up: This can be fairly nuanced, at this stage the user is eligible to convert again based on product usage, subscription term, or content release cycle and has not performed another converting action. If a user has an annual subscription, this period would start on the day of their renewal.
- Sunset: This is specifically users who have not converted within ~90 days of their Re-up window. For a more robust definition, you can incorporate engagement metrics here to determine if a user should continue to be interacted with.
Adjust these definitions based on the seasonality of your business; i.e., if you sell Christmas trees, your definitions might be 12 months long before you consider a user in the Re-Up period.
Retaining Users on their First Day
Lifecycle audience: New
First-day retention is often misunderstood and rife with opinions and conflict between product owners and marketers. The general rule of thumb is after the user first engages with your brand, their likelihood to re-engage with you declines rapidly over a 7-day period. There are numerous causes for this, such as they found a solution elsewhere, the price was too high and they decided to punt on the decision, or they simply forgot. The focus of this strategy is to combat users who forget/get too busy. However, these strategies can be adapted for other reasons based on the content you leverage.
Collect Permissions
A catch-all here, but really, retention efforts can only realistically start when you know who your audience is. Prioritize acquiring a push subscription, email, or phone number so they can be retargeted in owned channels. You also have the option of gating certain areas of the application, requiring some information from the user before proceeding. Bear in mind this is something you should test before rolling out; around 15% of the brands I’ve worked with have seen significantly higher conversion by avoiding this tactic. Other areas you can test are prompting sign-up as the user progresses through the app either through a session duration, key screens, or other browsing behavior.
What about users who never give me their information? This is a common problem. Not every user who downloads is going to give you a juicy morsel of retargeting goodness, so you’ll have to test out other ways to reach them—especially paid audiences. Experiment with soft prompts earlier on, such as collecting App Tracking Transparency (ATT for iOS) permissions to access IDFAs for retargeting, or prompting for foreground push notifications (iOS + Android). Brands that are successfully collecting permissions are typically using an In-App notification to prompt for permissions before requesting them directly. If you don’t have the means of retargeting users based on the information you’ve collected about them, consider a tool like AppsFlyer that enables capture and transmission to retargeting partners.
Message Fast
It’s in the name: First-Day Retention. Getting your message out there fast will have the biggest impact on the likelihood for users to retain, especially as they compare shops or put off the decision. Typically these messages are reserved for abandonment-type campaigns that generally have the largest success. You can also use these as opportunities to prompt users for additional info or clarification.
In terms of timing, within 1-2 hours is generally best, although I’ve seen success with brands messaging as early as 15 minutes.
A few examples of messages you can test out:
- Like [Product name]? Check out more [Product Category].
- Considering a [Product Name]? Tap here for tips on how to decide the best [Product Category] for you.
- Booking your appointment with Dr. Gordon? Availability is limited, reserve your spot now.
- Using action buttons to ask a question
In terms of timing, within 1-2 hours is generally best, although I’ve seen success with brands messaging as early as 15 minutes.
Retaining Users in Their First Week
Lifecycle audience: New
Immediacy is the key to keeping users retained in the app. The purpose of the first-week focus is to continue to leverage a combination of immediacy and relevance to create a messaging flow that keeps the user engaged to reach a decision on conversion.
Weave a Story
Generally speaking, your toolkit will largely lean on a welcome or nurture series. Reach out early and often — customer engagement typically declines each day. In addition, you can make these messages more dynamic and personalized based on behavioral cues you’re collecting, such as which specific items they are browsing, category they are looking at, screens viewed, step in the registration funnel they fell off, etc. Lastly, continue to collect more data about them, their preferences, and their thoughts by using a mix of push + SMS + in-app notifications.
Let’s take a look at a branching nurture series from one of our clients in the healthcare industry.
Now they leverage mobile-only channels, but this flow can be combined with email, retargeting, and other channels as well. Note that we’re looking at two specific behavioral cues: “Have they registered?” and “Have they booked?”.
There’s other information we can siphon based on their level of interaction: Which doctor or specialty they browsing, what symptoms have they described in their registration flow, etc. This kind of exercise can be expanded for additional use cases to surface other data points we may want to collect, in which we can leverage in our marketing program.
Progressive Personalization
Each time the user comes back during this period (and beyond), you can continue to increase the level of personalization in your content based on their browsing behavior. Let’s look at an example of these messages for the mental health offerings of our healthcare example. Below is a standard welcome / nurture series:
Message on Day 1:
Schedule your first appointment with a trusted therapist in [City]
Message on Day 3:
Testimonials from our clients
Message on Day 6:
Meet one of our therapists
Now let’s look at how we would modify this using incremental personalization based on addressing the potential patient’s concerns, either provided from feedback gathered via surveys or user studies, and weaving a story around that. Note that this tactic leverages information you have, so each message would have a more general fallback.
Message on Day 1 – Value Prop
Scheduling time in person with a therapist in [City name/region] is hard. Get trusted advice from a licensed therapist on your schedule.
Message on Day 2 – Outcomes
[Dr. Name browsed] has helped [dozens/hundreds] of individuals achieve their personal health outcomes, like [insert like-gender name / Beauregard | Agnus] who said “I struggled with anxiety for years and [Dr. Name] really helped me work out the root causes and gave me the tools to better myself
Message on Day 4 – Hesitation
Is therapy like being on another Zoom call? How long does it take? Our experts answer questions from our customers.
Building journeys like this can vary in difficulty depending on your toolkit. If you find this technically difficult to achieve, consider a tool like Iterable to enable personalization across your mobile campaigns at scale.
Winning Back Lost Users
Lifecycle audience: Lapsed / Re-Up
Typically users who have been sitting in your database and haven’t engaged in the last 30 days are unlikely to return at all, let alone 3 months. Set the expectation that regardless of how good your content is or how many channels you activate, winback rates from campaigns generally range from 0.5 – 2% of the total pool. It’s important that you rightsize your investment to the reality of how much impact you can have, especially on owned channels.
After users have gone through your Winback series, you will want to sunset them, essentially removing them from your core messaging audience until a future executive sees an opportunity in messaging them (marketing jokes, amirite?!).
Automate Quickly
Winback rates and revenue upside generally mean the effort for any manual campaign does not pay off. Even winback users in your “best campaigns” (especially in November for retailers) have generally just as much impact as a manual series you put together. In order to stave off needless effort that doesn’t result in rightsized revenue outcomes, focus on automating your winback campaigns first, then shift towards optimizing around that winback program.
Establish a Cadence
As mentioned earlier, that rule of degrading engagement over time holds up for winback, so it’s best to start your winback campaigns early and do them often. Depending on your brand and customer lifecycle, this means starting between day 30 to day 90 of their last engagement with you. Unlike days 1–7, generally, there is a less drastic impact of diminishing returns on each message, but it does level out fairly quickly. Your first effort may result in 0.3% of your audience won back, whereas subsequent efforts may yield 0.1% to 0.2% respectively—at some point, it just isn’t worth it to send anymore. A starting point for your winback program is to cap it around 4 touch points.
Leverage Multiple Channels
Some causes of winback are simply abandonment of engagement on that channel. Segment by their last engagement channel and try messaging in another format if you have it available. Just be mindful that the cost to send those messages (especially SMS) doesn’t exceed the revenue upside. In the example below for a meditation app, there is a natural progression of common challenges—why the user installed the app in the first place (default to your largest category if they never engaged meaningfully).
In this case, the user provided us enough signals to determine their challenge was related to sleep; the content shown here helps us get to the heart of the matter of whether the user is truly disengaged with the brand, deactivated on that channel, or simply found the solution elsewhere.
Experiment with different formats.
Winback audiences are a great area to experiment with different creative treatments, especially plain text formats that look like “letters from customer success”. This is a tried and true playbook tactic that results in great performance across many brands I’ve worked with.
As in the example above, you can also leverage native functionality in push and SMS messaging channels to ask questions to help re-engage the user. Consider the idea of having a user opt-out with a message like “Did you solve your issue without us?” as this can help reduce the cost of winning them back.
Wrapping Up
These examples showcase that there are various strategies you can employ to keep users engaged. To scale your efforts, use these examples as starting points to build a fully automated program. Even with limited data or marketing tooling capability, there’s still a lot you can do with variations in content, timing, and channels that will build a strong foundation. If you end up with 25 variations on your welcome series and you’ve validated the lift for each branch via A/B testing, that’s representative of what the best brands are doing today.
Steven Aldrich
Steven Aldrich is the Co-Founder and Co-CEO of Ragnarok, an agency that blends Marketing acumen with MarTech consultancy. For more than 15 years, Steven has helped brands improve their app retention through strategic mapping, increasing automation and personalization, implementing new marketing tools, and training teams to think more strategically.
