Precision Retention

We went to Fort Lauderdale when my kids were off school. We stayed one night, spent time in Miami, and came home. We loved the hotel we stayed in. It was in a perfect location, the room was great, and they had unique amenities. You could rent bikes for free or use the golf cart service to travel around the area.

But… we didn’t use any of the amenities.

And it wasn’t cheap. The price of everything is high and getting higher, but the hotel for one night didn’t feel worth it to me. For my wife it did. When we were talking about it, we were trying to decide if the value was fair.

The Value Equation: Output ≥ Inputs

We spent X dollars; we got Y in return.

X is easy to define. It’s a dollar amount. But how do you define Y? Each situation is unique.

For the hotel experience, our Y was:

• A safe and clean place to stay.
• Simple parking options for our car.
• Proximity to activities and restaurants.

The bikes, golf cart, restaurant in the hotel, and other amenities we didn’t use were part of the Y for the hotel. When they set the price, they consider the full package. For us, as customers, we consider Y relative to what we actually use. This is perceived value vs. realized value. And for us to realize the value, our actual Y must meet or exceed the actual X.

Now consider software products. Software as a Service (SaaS) is similar. You provide a service, people pay for the service, and they decide whether The Value Equation is:

1. Positive because the outputs meet or exceed the inputs—realized value.
2. Negative because the output is less than the input—perceived, not realized value.

Did I get value from it?

When it’s positive, the customer keeps paying you. And they’ll keep paying so long as the value stays positive. If that ever shifts, then you lose a customer. In the SaaS game, this is called “churn”. You need to add and retain more customers than you lose.

Churn is a problem on every software product I’ve worked on. Subscription fatigue is compounding. People are growing restless when there’s a monthly cost for everything.

It’s never a one-time calculation. Value must be met continuously to justify the spend.

Repeat customers are everything.

Why? Predictability.

If Sarah shows up to my coffee shop every day and spends $5, I can count on that revenue. The more she buys the coffee, the more she relies on it, the better I know her, and the more likely I can sell her muffins, croissants, or any other item my coffee shop creates because of the established trust.

This is expansion. You expand the revenue of people already paying for your product. You learn what’s working and do more of that instead of chasing what’s not working and obsess over every lost customer.

Retaining customers is hard. New software products pop up every day. Staying ahead of the curve and delivering new value is an infinite game. Technology always evolves.

What matters is how you focus on retention, understanding how and why people use your products. This includes why they stop using your product. And then you design the right interventions to make it better before you translate the signals into actions.

I call this Precision Retention.

The three pillars are:

1. Layer Diagnosis to find where value breaks down.
2. Intervention Design to create targeted experiments.
3. Signal Selection to focus on what actually matters.

Layer Diagnosis

The hotel amenities like the free bikes and the golf cart weren’t part of our value equation. We didn’t use them. Why? It’s not because we didn’t know about them or know how to use them. It’s because we didn’t feel the need or desire to rent the bikes or take advantage of the golf cart—unrealized value.

But before the hotel can diagnose the problem of unrealized value for a potential return customer, they need to know where in the process the value was lost.

I call this Layer Diagnosis:

1. Did they know about it?
2. Did they know how to use it?
3. Did they realize value from it?

Visibility + Comprehension + Realization

Find where value breaks down. Each layer requires different solutions. Only when you know where the value breaks down can you move into the solution space.

Adoption Physics is the macro lens for understanding why people adopt a solution.

Layer Diagnosis is how you identify the root cause in the value chain.

Where is the breakdown?

If you want to keep customers around, you need to know how to help them achieve continuous realized value.

Visibility

Let’s assume we took a survey after our stay in the hotel. They ask us about our stay and if we’d come back. They include questions about the amenities, with a specific option to notate whether we were aware of such amenities.

Let’s say we didn’t know about the amenities.

That’s a visibility problem. If we don’t know the amenities exist, we can’t expect to realize the value. The problem stops here.

This happens all the time in software.

Customers, unaware of features, leave the product for a competitor. They cite “budget” as the reason because it’s an easy scapegoat, but they probably just didn’t know your product could do exactly what they were looking for. You didn’t show them.

What’s the solution?

Raise awareness. Show up where your customer does: emails, texts, ads, social media, you name it. Multiple channels increases the likelihood you’ll be noticed. In a world of endless notifications pining for the attention of everyone, it’s not easy. It takes focused experimentation to find the right pathway, and diligence to keep on top of it.

For the hotel, we already opted into the experience. My wife knew about the amenities through exploring the website and the paper we received when we checked in. Software products do it through onboarding emails or in-app notifications. They show key features, point to demos or landing pages, and hit you with the barrage of “did you know about this feature?” messages.

Comprehension

Maybe they know about the product or feature or service. We did with the bikes and golf cart. But what if they don’t know how to use it?

Then it’s a comprehension problem.

Hotels have people who work at the front desk helping guests learn how to use the amenities provided by the hotel. Software relies on demos, emails, product messaging, webinars, and making a naturally intuitive product.

If people don’t know how to use a feature, they’ll never realize the value from it.

What’s the solution?

Close the comprehension gap. Find out why someone doesn’t know how to use the feature, and then iteratively make it better. Look at quantitative and qualitative data to see where the process towards value falls short. Use your intuition as the additional data point to shape the simplest version of the feature.

Less is always more. Simple is better.

If we didn’t know how to use the free bikes at the hotel, there’s a number of options the hotel could try to close the comprehension gap.

• Have a small printout in the room.
• Put a sign in the lobby.
• Tell you at check-in.

When it comes to software, multi-channel outreach helps. Show the information in the product, through emails, on social media, and even in webinar presentations. Know your audience, where they show up, and how you can best empower them with the knowledge they need to understand the value.

Realization

If they know about the feature and how to use it, the next challenge is the most difficult. And it’s also where most products struggle. The realization problem is when someone uses your feature, but doesn’t get value out of it.

The Value Equation is negative.

When assessing the value of the hotel stay, we were comparing the total value we felt against the total amount we paid. Value is an equation of inputs and outputs, comparing what we spend against what we receive. What’s tricky is our perception of value is widely varied person to person, moment to moment.

That’s why my wife felt the value was worth it when I did not. I enjoyed the experience, but also felt the cost was too high for it.

Realizing value is personal. That’s why you simply cannot accommodate everyone.

What’s the solution?

Get clear about the problem. If you don’t know what problem you’re solving, the value exchange won’t take place. There’s rarely one use-case for a feature, which is why product work often feels like threading the needle to solve the right problem in the simplest way.

Spend more time in the problem space. Go wide before you converge on one solution. Use your intuition to find the signals, and then solve iteratively in small increments.

Intervention Design

Diagnosing and understanding where the problem lies is the first step. Once you know which layer to focus on, then you design experiments to probe, to sense, to test.

Discussing churn, we explored the idea of updating product demos used in training.

The idea of tracking views and completions came up. I quickly pushed back. Building an entire system—or paying for one—to track video completions is a big lift. We don’t even know if the outdated or missing demos are part of the problem. Will people watch them? And how does it correlate with retention?

Layer Diagnosis draws the circle on the map.

Intervention Design walks the territory to understand the terrain, the weather, and the factors that affect the environment.

You know where to focus, but you don’t know the highest-leverage solution yet. And you don’t want to expend the energy on a solution without some degree of confidence in the outcome. That’s where experiments help.

What are we going to do about?

That’s the question. Depending on the layer, the solution should look different.

Visibility problems call for interventions to create awareness to move them forward.

Comprehension problems call for interventions to close the gap in understanding.

Realization problems call for interventions to find out what’s not working (and is working).

The solution to invest in training videos only makes sense when you know your audience can’t find the features and don’t know how to use them. And you need to be certain they can find demos, or you’re opening a new problem.

What do the interventions look like?

I define an intervention as:

• Name of the intervention.
• Trigger to initiate the intervention.
• Action to define what you actually do.
• Measurement to define what success is.

Let’s look at an example intervention for each of the layers. I’m keeping these generalized as interventions I use across each layer, but you should adjust them and experiment with your own interventions. These are meant to be fit to your specific needs and product.

Visibility Intervention

Name: Feature Shout-out

Trigger: A new feature’s usage is below the target usage rate in its first two weeks.

Action: Show an in-product notification when the user is near the feature and link them to a short demo showcasing the feature.

Measurement: The usage rate meets or exceeds the target in the next two weeks.

Comprehension Intervention

Name: Blank Slate Example

Trigger: A large percentage of users visit a feature multiple times without taking action.

Action: Create a preloaded example template showing what they can create with it.

Measurement: The increase of actions after the template is shown.

Realization Intervention

Name: ‘Almost There!’ Nudge

Trigger: User generates an output, then takes no downstream action within T (e.g., 10 min).

Action: Email that frames the next step as small and inevitable: “Your output is ready—finish the last step to get the value.”

Measurement: Downstream action rate increases for the cohort over T.

Intervention Library

These are just a few examples. Once you start creating targeted interventions, you can build a library of interventions to reuse and refine over time. I added a set of interventions to a Notion page where you can view them and repurpose as needed: Product Interventions.

Signal Selection

Not all problems need solving. There’s only so much energy your team can commit to solving problems. And problems in software are wide-reaching and multi-dimensional.

Knowing that, you need to focus on what really matters—the signal.

What’s failing? What’s working? What’s noise?

An even better question: Should we even fix this?

Take the hotel example again. We have young kids. That puts us in a specific demographic segment of customers. Families traveling have different needs than someone on a business trip or a couple on a romantic getaway. Hotels, like software products, serve a range of users.

And you can’t solve the problem for everyone.

The common advice in software is to narrow your target users so you know how to reach them, how to solve their problems, and how to create value so they keep showing up. The nice restaurant at the hotel isn’t worth promoting to us because we wouldn’t subject everyone to a dinner experience with our kids. So then the hotel shouldn’t fret when we don’t show up there. It was never part of our value equation.

Signal Selection is using the lessons from the targeted interventions to create focus.

Visibility problem? Focus your energy on creating awareness.

Comprehension problem? Focus your energy on creating understanding.

Realization problem? Focus your energy on creating value.

Like I said at the outset, don’t try to solve every problem.

There are two angles to consider:

1. Natural Attrition: If they don’t get value out of your solution, you need to improve it or accept the natural attrition. People stop using products, stay at different hotels, buy different products. Chasing everyone down is a fruitless endeavor.
2. Negative Inversion: Don’t just focus on what’s not working; learn from what is working. This is especially true when it comes to retention. I watch a lot of teams focus exclusively on the problems instead of the highlights, the opportunities, the bright spots. Then use them to filter the noise and focus on what works.

Finding a signal requires attention, intention, and focus. You solve problems more effectively when exerting energy in one area, like the sun focused through a magnifying glass.

The Practice

The practice of Precision Retention in five steps:

1. Diagnose: Where is value breaking down? Pick one cohort and one behavior to classify where the breakdown is: visibility, comprehension, realization.
2. Design: Create a single Intervention Spec with name, trigger, action, measurement.
3. Deploy: Run the intervention test on a small, controlled cohort.
4. Decide: After a the established measurement window, evaluate. If yes, the signal exists; if no, it’s likely the wrong layer or a weak action you should redesign.
5. Expand: If you see a signal, automate the trigger, scale the action, track the metric. If you don’t see a signal, adjust the layer, redesign the intervention, or drop it entirely.

Find one breakdown, test one intervention, learn what matters. And then repeat.

Use the Product Interventions as example starting points.

The Throughline

Delivering value is nuanced. There’s no one right way. The philosophy of Equilio is centered on the ideas of intuition, integration, and iteration. And I continue to see these three elements permeate the work. They’ve become a personal philosophy of craft as I navigate expanding my own working intelligence alongside AI and my team.

Intuition comes from paying attention, being mindful, and focusing your energy.

Integration comes from diverse knowledge and the translation of information.

Iteration comes from focused bursts of action to probe, to sense, to learn.

Together, they imbue Precision Retention, and all the concepts I talk about, with forward motion through continuous improvement.

If you want to deliver value and retain customers, you need intuition to know which layer to focus on, integration to design interventions in a complex system, and iteration to keep trying new methods to refine quality and solve your customer’s problems.

Connected Ideas

Precision Retention lives in the Clarity Current of the Claritorium and Quality Refinement of Equilio. And it connects to other ideas:

• Adoption Physics → Where breakdowns happen
• Experience Field → Why they happen
• The Fieldbook → How you test fixes
• Signal Practice → How you learn
• Attention Allocation → What you prioritize