Building a sustainable business in the product era means working with a new set of assumptions and goals. So far in this playbook, we’ve covered how to think about product engagement using the three games of engagement model, how to activate new users, and how to re-engage users with engagement loops and triggers. To achieve any of this, however, you need to have a true modern engagement stack in place.
In order to win, businesses need a modern engagement stack because product development has fundamentally changed.
In the past, software companies could ask customers to pay an up-front fee for a product, then roll out updates every once in a while, and ask those customers to buy the new version on a yearly basis (sometimes even longer!).
Now, in the digital age, both the nature and the speed at which we build products—and, consequently, the way we grow and monetize businesses—has fundamentally changed. With the exploding number of product offerings in every market, the companies that make it big are the ones that know how to grow by meeting the changing demands of their customers and continuously delivering value.
*In 2005, “online” was a new acquisition channel, now it’s the product.”
Product analytics, which we’ve highlighted in every chapter of this playbook, is powerful for analyzing in-product engagement and deciding what to build next.
But to really drive your North Star metric and execute on a full engagement strategy that aligns across product and marketing, you need an engagement stack:
best-in-class software to help you get data into your analytics solution, test hypotheses, assign attribution, do marketing automation, and more.
Why do we support the best-in-breed software approach versus other alternatives like building in-house or an all-in-one solution? Well, it is impossible to find a third-party all-in-one solution for engagement that is truly excellent. And in-house tools are often hacked together and/or complicated to use, making them only usable for the most technical teams within a company. Moreover, they’re usually slow, difficult to maintain, and generally inflexible.
Instead, we believe every business should build an engagement stack suited to their needs using the best-in-breed tools that are available for each layer.
To develop the next several sections of this chapter, we had the pleasure of speaking with Austin Hay. In his role as VP of Consulting at The Growth Practice, a subsidiary of mParticle, and GM of HBE Ventures, he helps Fortune 500 businesses and enterprises like Walmart, Turner, and Sam’s Club evaluate, implement, and maintain their growth stack.
When you’re figuring out exactly which combination of tools should make up your engagement stack, you need to consider two things:
1. What you want to accomplish from a business perspective
2. How you want to prioritize the different initiatives you want to take on
Says Hay, “The number one thing to do is just be very objective and straightforward in outlining your business criteria upfront. Go through and document, ‘I want to be able to achieve X. I want to be able to create an audience. I want the capacity to have a UI that sends data as an output to Y.’”
Knowing exactly what your business objectives are will then help you prioritize Knowing exactly what your business objectives are will then help you prioritize your initiatives. your initiatives. This, in turn, will help you understand the tools and systems you need in place to be successful.
Here are some general best practices for setting up an engagement stack. We’ll be referring to some of these in more detail later in the chapter.
|Choose tools that do one or a couple of things really really well.||Go for solutions that promise to do a hundred things mediocrely.|
|Choose tools that allow you to move data in and out easily, thus enabling interoperation with other engagement software.||Choose tools that are closed off and can’t work in an ecosystem with other tools.|
|Choose tools that democratize data for all teams.||Choose tools that require endless configuration and hacking to get work done.|
|Consider choosing tools that are known to be reliable and built by experienced teams.||Jump to using the coolest new technology without proper vetting—you don’t want to risk an integral layer of your engagement stack suddenly going out of business.|
|Think through the order of implementation for your stack.||Drink a seller’s Kool-Aid and jump to implementing the most interesting tool.|
|Document your stack, including why you implemented certain tools and when/why you make changes to the stack.|
Every company’s needs are different, so are the tools they’ll need to measure and drive engagement. Here are the main layers we believe any product-led business should consider when building their engagement stack. There are many tools for each of these categories; we’ve included some of the best in the industry that work well with Amplitude.
Web and mobile applications generate immense volumes of customer data. For this data to have any value, teams should be able to better understand their customers and then act on those insights with both product and marketing efforts.
Customer data platforms make this easy to do by pulling data from multiple sources, then cleaning and consolidating it into a single database. Using a CDP is the simplest way to connect all the tools in your engagement stack—you can easily send customer data to your analytics platform for analysis, and also to the experimentation and marketing layers of your stack when it’s time to act on the insights.
Customer data platforms let teams focus on building better experiences for their customers, instead of worrying about how to get data from tool to tool.
If you’re reading this playbook, then you probably already have some idea of the value of product analytics. Building an engagement stack requires that you have not just data, but the means to get the insights that help you build an engaging product.
As we covered in great depth in the previous chapters, product analytics allows teams to ask deep questions of their customers’ behavior. Questions like:
– How long does it take a new user to become a habitual user?
– What kind of user personas do I have in my product and how do they behave differently?
– What actions do users who convert do differently compared to users who do not?
– How can I reactivate lapsed users and get them to see the value in the product?
– How has our new marketing campaign or product launch changed user growth in the last 3 months?
Answers to questions like these lead to ideas for product improvements, new experiences, or campaigns that are backed by data. Dashboards with vanity metrics like page views or app downloads are not enough.
Another key value proposition of a product analytics platform like Amplitude is unifying different data streams into a single source of truth. Not only do businesses care about bringing user data into their systems, but also sending data out to marketing vendors.
When we asked Austin Hay about why analytics should be a core part of anyone’s engagement stack, he said, “It’s a two-fold thing. It’s [the ability to understand] one, what marketing-based behaviors and actions can we take off this funnel, and two, what qualities about the product experience can we understand and then make better so that we can have a better experience that requires less activation energy.”
The attribution layer fills in the context of a user when they first get into your product. The reason this is more complicated than it sounds is simply because the user journey is more complicated than it used to be. People can come into your product through a variety of ways—referrals, ads, App Store searches, and so on.
Attribution is particularly important if you plan to run and measure the impact of paid campaigns. If you don’t have a proper linking infrastructure in place before
you start measuring engagement, you can’t be confident that the data you’re looking at is sound and complete.
Attribution tools that can also do deep linking are useful for actually leveraging that data to point users to specific content on the web or in a mobile app.
An attribution layer lets you immediately assess how a user arrived in your product, where they came from, and all of the other information that can help you distinguish them, even when the path into your product is a complex, multi-channel journey.
Experimentation starts with building conviction around an idea before spending significant resources to roll it out to all of your users. The most reliable way
to execute on your ideas for improvement is to A/B test them.
In an A/B test, you deliver two experiences to users. One features your product in its original state; the other features a version of your product with some change. After a certain period of time elapses, you can examine the results of the experiment to see if there was a significant difference between the two groups.
In the previous chapters, we talked about how to measure and build for in-product engagement. A good engagement stack also includes a marketing layer to engage users outside of the product.
Whether through email or in-app messaging, push notification or SMS, marketing helps you ask users for feedback, provide them with relevant content and nudge them in the right direction.
The information you gather from all of your product’s digital touchpoints is crucial to understanding how you should be re-engaging customers. For example, if you figure out that some segment of users aren’t activating in your product, you can send them a personalized message to ask for more information. Given what you learn from that request, you might decide on any number of ways to grow awareness and understanding of the new feature among your wider user base—from a customized transactional email campaign to a well-timed push notification.
Many of the “external triggers” that we mentioned in Chapter 3 as tactics for re-engagement can be delivered through marketing automation tools.
– Urban Airship
One of the big mistakes that teams make in building their growth stack, according to Hay, is not prioritizing which layers of your stack to implement first.
Hay suggests starting with a customer data platform first, which will make Tip: Start with a customer data platform first. integration with product analytics (we recommend Amplitude) easy. Once you have your customer data platform and your analytics layer, then focus on what tools you’ll need to carry out your next most important business objectives.
Once you decide on the layers of your engagement stack and finish implementing them, the next step is to gain adoption within your organization.
Adoption is critical. The quicker you can implement and understand a tool, the The quicker you can implement and understand a tool, the quicker you and others in your org can get value. quicker you and others in your org can get value. Getting repeat value from the tool ensures that it becomes a true fixture in your daily work. “If not, you typically end up churning, looking for other tools, repeating, and before you know it, you’ve gone through three or four tools in the course of a couple years,” says Hay.
According to David Reyneke, Director of Growth at Prolific Interactive, a mobile-focused product development agency: Repetition is the key to getting value out of your engagement stack.
Reyneke suggests a simple framework for getting adoption:
– Start by performing a sprint-like runthrough of all the tools in your stack.
– Allow people to play around with the more intuitive tools to spark curiosity.
– Set parameters and goals around usage of the tools for the first few weeks.
Reyneke describes one example workflow, “For Braze and Amplitude, we would recommend every week sitting in Amplitude, looking for insights, and setting a KPI that you want to try to move the needle on for the week. Then, you execute that on that in Braze over the course of the week with experiments. Then you evaluate your progress in Amplitude.”
Having at least one team member work in multiple tools regularly will quickly demonstrate how each part of the stack works.
Once you have the right tools in your stack and all the right people in your organization using them (easier said than done, we know), the third thing you’ll need to keep in mind is maintenance.
As your business scales, it’s important to periodically go through and audit your stack. More often than not, as people leave, find other tools, and time passes, maintenance winds up being a thankless task that falls through the cracks.
It is best to start documenting early and make sure there’s a process in place to keep your docs updated. Some tips:
– Keep nomenclature consistent by creating a spec sheet for all the data points you want to capture at the very beginning.
– Include in your spec sheet: what tools you’re using; how and why they were implemented; which tools are sending data where.
– Assign someone to maintain the spec sheet.
– Update the spec sheet every time a change is made to the stack.
“How you implemented things, why you implemented things, when you make changes, why you make changes: all of that should live in a document that lives with the company,” says Hay.
In the final chapter of this playbook, we discussed why you need an engagement stack and our recommendations for the specific tools to choose for your business: a customer data platform, product analytics, attribution and deep linking, A/B testing software, and marking automation. We went over some best practices for setting up your stack, gaining internal adoption, and maintaining your stack as your business scales.
Take a moment now to reflect on your learnings. Consider the following:
– What does your engagement stack look like today? Is there up-to-date documentation on it?
– What does each tool in your stack do? What data flows in/out?
– How many people on your team or organization have adopted these tools?
– Are able to get a full picture of your customer journey and act on those insights using your engagement stack?
– Is your stack meeting the needs of your business?
How great product companies are transforming their product analytics stack
Justin Bauer, VP Product at Amplitude
Austin Hay, VP Consulting Services at The Growth Practice
Building Apps for Mobile Growth: Tech Layer Choices
Andy Carvell, Co-founder at Phiture Mobile Growth Consultancy
We hope this installment of the Product Analytics Playbook series gave you a good foundation for building products that engage and deliver value to your users. We wrote this to help you measure and improve product engagement through best practices, frameworks, and models that we believe in and use ourselves, here at Amplitude.
How are you using this content to build better products in your own team or organization? We’d love to hear from you with questions, comments, epiphanies, or just a ‘hello’. Email us at email@example.com.
Product Analytics Playbook I: Mastering Retention
Learn proven methods for building a data-informed retention strategy.
Get it at productanalyticsplaybook.com