Raise prices, don’t get too hungry to eat. Ref: Marc Andreessen in Tools of Titans. 

Monetization Challenges

  • Sacred Cows: successful products are afraid of changing their monetization model since it has worked so far.
  • Fear of customer revolt: the longer you go without changing pricing the more that a customer revolt increases.
  • Ripple effects on Growth: monetization needs to account for the full growth loops.
  • Lots of stakeholders slows things down: Leadership, Finance, CS, Sales, Acquisition & Retention, etc.
  • Infrastructure lift: Contracts, FAQs, Support docs, Pricing docs, Finance pricing, etc.

Monetization is a barrier for both you and your competition, so working on it puts you at an advantage if you do it well. Profitwell published some data showing that improvements in monetization yielded a 2x to 4x greater impact on the bottom line than improvements to acquisition and retention.

Monetization Mistakes and Principles


  1. Mistake: Pricing only with the business view. Instead, you must align monetization strategy with value as the consumer or customer perceives it. (e.g., SurveyMonkey Enterprise plan assumed that Admin and Security features were very valuable, in reality the product was adopted separately by smaller teams – had to repackage it.)
  2. Mistake: Monetization siloed from growth. Pricing impacts your growth loop models, so that needs to be factored in (e.g., Zoom free tier leads to viral growth, Airbnb had no hosts fee when they needed hosts for growth).
  3. Mistake: Consider all revenue as good. You need to consider both acquisition cost and cost to serve when pricing components of your product, so you can account for contribution margin and payback periods.
  4. Mistake: Only consider price. Priced is zero-sum, the incremental dollar is taken from the customer. Focus on other areas: use cases to monetize, features or attributes to monetize, price scaling, amount, and timing of charge.
  5. Mistake: Set-and-forget monetization strategy. Need to adapt as the business changes. (e.g., Airbnb charge to hosts once enough hosts on platform, Netflix changed structure when streaming became a different focus and cost structure).
  6. Mistake: Use monetization as a short-term growth lever. Need to consider long-term growth impact. (e.g., Discounting can help acquire more customers, but they may be lower quality customers and value may erode if discounts become expected.)

Use Case Model

  1. Problem: what problem are you solving for, in the words of the user. The words of the user reflect both the problem and the persona articulating it.
  2. Persona: who are you solving it for? Monetization needs to be based on the characteristics of the persona you provide value to.
  3. Alternative: alternatives will be different per use case or plan. Consider what they’d do or use if your product doesn’t exist.
  4. Why: why they choose you vs. alternative? Motivation to use our product (personal utility, financial, social) and how it’s different than alternative (all-in-one tool, most personalized, selection/speed, etc.).
  5. Frequency: natural frequency influences pricing based on expected interaction frequency.

Monetization Model

Things to understand

  1. How do price scales with value?
    • Feature Differentiated: price scales to add features
    • Usage Value Metric: function of usage
    • Outcome value metric: per outcome such as ticket sale, order, etc.
    • Continuous (per user/item/etc.) vs. Banded (first 10 = X, next 10 is Y, etc.) — could be mixed.
  2. What features or attributes customers get from each use case?
  3. How much do we charge for the what, per customer per year? (think per customer, not per user)
  4. When do we charge for the what? (never, per transaction/use, monthly, yearly, every few years)

Model Friction

Lower friction makes it easier to convert.

  1. Scale: When value metrics are easy to measure and predict -> low friction. The harder to predict more friction. Outcome-based is less friction than usage based.
  2. What: is it easy to understand what features/offering I’m being charged for? (e.g., ClassPass harder to understand at first than a traditional gym membership)
  3. Amount: lower cost, lower friction
  4. When: never or high frequency (e.g., monthly) lower friction than long-term commitments.

Need to understand where does my model increase and decrease friction. For example, Slack has higher friction when going to Enterprise plan — annual upfront commitment.

Monetization Outputs

  1. How do I measure the output of my monetization strategy?
    • Top line revenue: Bookings, Revenue, Gross Merchandise Value or GMV – total transaction value on platform
    • Profitability: Margins, Net Contributions Margins (considers cost to acquire), Unit Economics (per unit)
  2. Revenue vs. Profitability
    • Top line revenue is critical to build a large company – typically 100% YoY growth when under $100M in yearly revenue, then ~50% YoY after that.
    • Profitability is more important to prioritize earlier for low-margin businesses.
    • Growth loops: sometimes funding upfront with low profitability can be key to drive growth loops (e.g., Airbnb needed to invest in acquiring hosts/users first and then could build profitability, Netflix needed to invest in content)

Analyze Revenue

  1. Choose a frequency – daily, weekly or monthly. Choose based on:
    • what questions to answer (e.g., daily campaigns) or investment decisions (monthly, quarterly or yearly).
    • audience
  2. Visualize revenue as a time series: look at trends, and growth rate as % (accelerating or decelerating?) Consider seasonality when choosing time frames. Break it down by use case / plan in both absolute dollars and growth rates.
  3. Revenue equation: revenue = breadth (number of paying customers) x depth (total revenue per customer). Look at the key variables for both breadth and depth. Plot and review the revenue equation components in time series to see how they are trending. Consider looking at them for each use case and business segment.
  4. Segment new vs. repeat: helps us understand health (repeat shows quality usage), consider monetization models (e.g., monthly billing converts faster but retains less than annual), and optimization strategies (new vs. existing customers may need different incentives, etc.)
    • Segment your revenue and revenue equation components segmented by new vs. repeat.

Revenue Creation Cohorts

Building NEW Revenue Creation Cohorts

  1. For New Revenue, need to start looking at user from start point before new revenue. What leads to revenue. (e.g., for Zoom you look at free user signup – not when they pay for a plan)
  2. Time Period: match frequency based on hypothesis on how long it takes someone from initials sign up to transaction.
  3. Revenue creation (how much revenue a cohort generates and how quickly?) or customer conversion (how many paying customers per cohort)
    • Can look a cumulative vs. non-cumulative.
    • To compare customer conversion across cohorts, use %ages to normalize by cohort size.
    • To compare revenue per account across cohort, look at average revenue (normalize by customers).

Analyze NEW Revenue Creation Cohorts

  1. High-level: customer conversion velocity (plot avg of customer conversion cohorts); revenue conversion velocity (do the same with revenue – can help us understand the average value of a lead).
  2. Individual cohorts: drill down to look at whether things are improving or not (cohort size, acceleration/deceleration, outliers) by plotting each cohort separately. Look at a heat map of the cohort tables to identify trends. Diagonal trends would be when an issue impacted all cohorts at the same time (e.g., a bug).
  3. Segments: acquisition parameters, persona attributes, user actions in lifecycle. e.g., by team size, by category

Building Revenue Retention Cohorts

Revenue retention is key to grow the business, otherwise you need to acquire even more new revenue to grow.

CAC increases over time due to: audience saturation, increased competition, moving from high-intent to low intent.

You need to answer:

  1. How is our monetization model retaining repeat revenue?
  2. Where is repeat revenue coming from (segment, time period)?

To answer these, build revenue retention cohorts. Decide on who is in the bucket, how much new revenue the cohort started with, a frequency to review the cohort, and non-cumulative revenue per time period. Use % of original revenue to compare cohort revenue retention.

Analyzing Revenue Retention Cohorts

Look at time-based net revenue retention (see how much revenue you retain), time-based customer retention (how many customers you retain), and time-based ARPC (avg rev per customer).

A good net retention rate varies by company strategy – do they land and expand? or do they grow by adding new customers? Consider regular market turnover.

Customer retention trends help us understand what will happen with net retention in the future, since you can see whether customers are dropping off and you rely on a smaller set of customers growing.

For ARPC you change the revenue by period to cumulative, instead of non-cumulative, and divide by cohort customer size.

After you’ve looked at the above overall, compare each cohort separately to see if retention is improving or not. Also, consider looking at segments separately particularly by acquisition method to highlight different revenue retention characteristics.

Analyzing Repeat Revenue States

States to consider:

  1. Existing: transacted in last time period.
    • Expansion: transacted in time period before, spent more in this time period. (added more users, upgraded tiers, etc.)
    • Contraction: transacted in time period before, spent less in this time period. (removed users, downgraded tiers, etc.)
  2. Churned: transacted in time period before, but nothing in this time period.
  3. Resurrection: transacted in the past, zero in time period before, spent something in this current time period.

To analyze your revenue states define the various stages (expansion, contraction, churn, resurrection).

You can use growth accounting bar and line charts to see how each category trends over time. You can also look at individual cohorts and compare by adding rows for each type of revenue state for each cohort. Also segment out by acquisition parameters, personas, actions to find trends or differences.

Cost of Revenue

Need consider cost in order to determine “net capital”, and use that to determine revenue quality. This net capital is then reinvested.

Many companies consider LTV:CAC, but it has problems:

  1. Assumptions that are difficult to get right: need lots of time and data to get it right. Average isn’t super helpful. Discounted cash flow methods help but not all.
  2. Doesn’t measure payback period – how fast money is reinvested in growth loops.
  3. Does not factor cost to serve and retain customers. Then cost to acquire.

Instead, let’s focus on a few things:

  1. Cost to Serve
    • Need to consider: physical products, logistics, product dev and maintenance, storage & hosting, customer support, program & tooling (CRMs, etc.), Partnerships & Integrations.
    • Ways costs can scale: variable, semi-variable, non-variable (usually avoid the non-variable costs).
    • Margin = revenue – cost to serve
    • Look at margin over time:
      • Plot revenue over time
      • Plot cost to serve over time
      • Calculate dollar margin over time (rev – cost to serve)
      • Calculate %ages.
    • You can then look at margin trends by use case.
  2. Cost to acquire
    • CAC: total cost (variable and semi variable) to acquire a customer. Include free trial cost to serve or for freemium customers include cost to serve free users.
      • Include: Ad costs, referral costs, people cost (e.g., sales reps, mktg), tooling & program costs (e.g., mailchimp), misc marketing costs.
      • Example: Figma is freemium, so all costs to support free tier is included in CAC. Figma would include: Ad campaign, inside sales, cost to serve free, cost to convert free to paid.
    • Net contribution margin = margin – cost to acquire
    • Payback Period: build cohorts and look at:
      • revenue creation cohorts
      • cost to serve by cohort – looks at avg cost to serve in a week, and multiply by number of customers in that cohort
      • then look at margin contribution by week (dollar margin)
      • Calculate cumulative dollar margins
      • add acquisition costs for the cohort
      • calculate net contribution margin and payback period (week when net contribution margin turns positive) for the cohort.
    • Look at cohorts over time to see if there are trends (e.g., lower payback periods in specific time periods)
    • Look at segmentation (e.g., acquisition parameters, persona, usage)

Model Opportunities

Consider the monetization triad: consumer view, growth loops and cost of revenue.

  1. Align the 3 for max revenue.
  2. Needs to evolve over time as Product (impacts all 3 – consumer view of value, growth loops and cost of revenue), Market, Audience, and Business evolve.

Consumer View: how value scale for the customer, what features and benefits they value, how much they are willing to pay, when/frequency they want to pay?

Growth Loops: consider gaps in growth loops when looking at where to add pricing.

Cost of Revenue: gaps between model and cost of revenue – if cost of revenue changes, need to re-align things.

Value metrics

Understand the consumer view first (as above, how value scales for them, what features are important, etc.)

  1. Feature differentiated, usage value metric, or outcome value metric.
  2. Each metric has pros/cons w.r.t. predictability of costs for consumers vs. alignment with core motivation.

To better understand the consumer view, you can do a pricing study and analyze it.

  1. Pricing surveys to analyze the value metric:
    • Ranking surveys: customers gives each feature a score based on importance to them.
    • Max-diff surveys: each customer chooses what they value most or least.
    • Conjoint analysis: give different options and let the customer pick one. Simulates actual behavior, but requires a high setup and analysis.
  2. When running a survey:
    • Define the scope: Why (motivations – personal, financial, social).
    • Only give ~3-6 dimensions, otherwise too difficult to rank. If more than that, break it into multiple sets.
  3. To analyze:
    • For max-diff, you can look at the relative difference score: (# of times chosen as MOST – # of times chosen as LEAST) / (# of times attribute shown in set)
    • Visualize score by plotting in a graph.
    • Segment by use cases or user persona, demographics, geography, etc.

With the above, you get a sense of what consumers value more from your product and what metrics they care about.

Feature value

A few considerations:

  1. Not all features are valued by users.
  2. Not all users/customers value the same things.
  3. Value does not equal willingness to pay – there may be alternatives or conditioning why we are not willing to pay for something.

You can do the max-diff survey on features as well.

  1. Define the scope: various value props or benefits, product features or attributes that users value.
  2. Survey users: max-diff as above (value metrics).
  3. Analyze: same as above (value metrics).

Willingness to pay

How much are users willing to pay for something? Their willingness to pay should be higher than the price. We should understand what to charge for, when, to what user segment, etc.

Consider: Use cases, Persona attribute, Usage

Van Westendorp study (easy to do, doesn’t model actual behavior) —

  1. ask the following questions: At what price point does the product become (please insert a dollar value only)
    • too expensive that you’d never consider purchasing it?
    • starting to become expensive, but you’d still consider purchasing it
    • a really good deal
    • too cheap that you’d question the quality of it
  2. Build the survey: describe the product, define the what (what are we asking them to pay for – key features), willingness to pay question (too expensive, not a bargain, not expensive, too cheap). Be specific as to what you’re asking them to buy, on what unit, for what time frame, etc.
  3. Plot things: do a cumulative chart, with price increasing in x axis, and y axis showing % of respondents. Cumulative would start a 0% on $0 for the too expensive and not a bargain categories, but would start at 100% and go down for the too cheap and not expensive categories.
  4. Analyze: look at the range when things are between not a bargain and too expensive. Also look at “optimal price point”, where too cheap and too expensive intersect. Use ranges to inform your pricing strategies. Don’t just pick a price.
  5. Segment results! Look at the ranges and data for various segments.

Conjoint analysis: simulates actual behavior, but incomplete picture of how users evaluate — only understand best/worst.

Live testing: simulates actual behavior, need significant volume and sophisticated infrastructure to handle and resolve various types of conflicts.


When do we charge? Look at customer natural frequency (never, per transaction, monthly, quarterly, yearly, multi-year)

First look at Willingness to Pay. Look at natural frequency and variance in frequency (e.g., Lyft may be low natural frequency when traveling but use a lot during travel). Also look at how often they want to make a decision on what products to use. Also look at how long it takes to build a habit around a value prop.

Transaction vs. Recurring:

  1. Look at Natural frequency and variance in natural frequency
  2. Qualitative: Ask users questions on frequency (how often do you face this problem, how often do you use the alternative)
  3. Quant: look at data by frequency of use
    • Frequency histogram (e.g., days active in the last 28 days)
    • Look at action and frequency of habit. (e.g., how many users ordered at least once per week in the 3 weeks).
    • Look at time to habit by segments of users. If it takes a while to establish a habit, a longer term contract may make sense.

Model Strategies

Drive more revenue by increasing breadth or depth.

Revenue = breadth x depth

Change an existing use case:

  1. Change the value metric: e.g., move from charging per view to charging per conversion.
  2. Change packaging
  3. Change pricing
  4. Add features to a plan

Expand use cases:

  1. Add a new tier
  2. Add more offerings (e.g., Uber adding Uber XL)

Change existing use cases when:

  1. Consumer doesn’t value it as we expected
  2. Growth loops aren’t enabled
  3. Cost of serving not balanced with revenue

Introduce a new use case:

  1. Vertically: something to add INSTEAD of existing. e.g., adding a new tier/plan.
  2. Horizontally: AND decision, such as an add-on.
  3. When?
    • Some segments have different view
    • Growth loops for some segments not enabled by hypothesis
    • Cost of serving SOME segments higher than revenue
    • Expand target audience and serve new personas/problems

Value Metric Strategy

You can have value based metrics based on Outcome, Usage, or Feature Differentiation. See more “value metric” information above.

Decide on: outcome vs. usage, continuous vs. banded, single vs. multiple.

Need to tie it back to the monetization triad: consumer value (see above max-diff survey), growth loops,

Calculate the scale willingness to pay:

  1. For a single respondent in Van Westendorp, take the avg value between “not a bargain” and “not expensive” values for a given value metric, then multiply that by the number of consumption units that they expect to pay for, which then gives you the “scaled willingness to pay” (e.g., a user says they’d be willing to pay between $100 to $200 per event and about 3 events per month gives you a scaled willingness to pay ~$450/mo)
  2. Visualize on a graph: scaled willingness to pay vs. number of events/value metric.
  3. Analyze: how well does willingness to pay scale with increase in value metric.
  4. Segment to see willingness of pay scale by group.

Issues when selecting the wrong value metric or how we charge:

  1. Growth loops: if we charge per user, then that will add friction to inviting another user which will harm a growth loop that requires inviting users.
  2. Cost structure: need to consider the cost to serve

Packaging Strategy

  1. Add features to existing packages
  2. Move features from one package to another

To define your packaging strategy:

  1. You first determine the relative preference score (max-diff on features as described above). Then segment the results by different customers personas.
  2. Look at willingness to pay for those features that customer personas really value.
  3. Plot the relative preference score vs. deviation from median willingness to pay.
    • Low/Low quadrant: important for setup or activation, business values it but customers don’t.
    • High value/low pay willingness: table stakes.
      • Drive expansion
    • High/High: great opportunities to package and price.
    • High willingness to pay/low value: not a lot of users value, but those that do are willing to pay a lot for it (e.g., customer success manager, etc.)
      • Consider these as add-ons

Pricing strategies

How to do good pricing changes:

  1. Align pricing with monetization triad: align with consumer view, growth loops, or cost of revenue.
  2. Align with other monetization model elements: e.g., new features, new value metric, etc.

To set your price:

  1. Van Westendorp: qualitative survey of willingness to pay that yields an acceptable range of prices.
  2. Conjoint: simulates actual buying behavior, which gives more optimal price point.
  3. Live testing: quantitative validation of optimal price point.

Consider overall impact of the price change:

  1. Typically an increase in price reduces conversion
  2. But it also yields extra profit that can be reinvested in growth loops such as Ads
  3. Need to balance the full equation to understand what will drive long term success.

When strategies

  1. Free vs. Paid
  2. Transactional vs. Subscription
  3. Term of subscription
    • Longer terms increases model friction, but increases retention. The longer a customer is committed the more likely that they’ll spend more time on the product and get to aha moments.
    • Look at how we incur cost – if upfront costs then more transactional is good. Also, if we incur a lot of upfront cost, longer subscription term is good.

Adding new use cases

  1. Vertical use cases
  2. Horizontal – things like add-ons that can be added to an existing tier/plan/etc.

To determine whether vertical or horizontal, look at:

  1. Which segments value the use case? If users across every vertical/existing plans value the new use case, an add-on/horizontal thing may work better.
  2. How does value scale for the new use case? If it scales the same way as existing use cases, vertical. If it scales differently, then horizontal (add-on) better.

Cost to serve

  1. An add-on/horizontal use case helps with large costs since it can be priced to only have friction for those with high willingness to pay.

Customer states

The goal of what you want to achieve with customers changes based on the customer state. Need to:

  1. Identify the customer needs
  2. Understand if they are likely to convert
  3. Then target them for conversion

Need to define various existing states:

  1. Potential customers
    • Goal to build new revenue
    • Get them to pay for the product and establish the habit.
  2. Healthy customers
  3. At-risk customers
    • Goal to maintain existing revenue
  4. Churned customers
    • Goal to bring back to existing revenue

You need to define target states to ensure who we are targeting and what we need to do.

Healthy state customers

  • Define:
    • Customer engages with the product based on the natural frequency of usage that we’ve defined.
    • Customer takes core actions that demonstrate value on that natural frequency (e.g., we see that core action happening)
    • Deepen revenue from current use cases (more frequency, more value per order, etc.)
    • Move customer to a higher ARPC use case (uptier, etc.)
    • Add on more use cases

To identify expansion candidates, look at healthy state customers and identify the target persona for expansion. Define the target persona, segment the data, and build a predictive model.

  1. Target persona
    • Who: customer persona attributes such as age, gender, household income, etc. Can also include the “how” we acquired them as an attribute.
    • What: actions taken in the product such as engagement, feature usage, hitting a paywall, etc.
    • When: external time-based triggers such as seasonal events or time of the year.
  2. Segmentation
    • Look at expansion rate by user attributes and/or action attributes, this gives you important triggers for expansion.
    • Then look for ideal expansion path profiles and lookalikes (those similar).
  3. Predictive model
    • Looks at attributes (who/what/when) and then using data science predict highest influence triggers and highest probability expansion candidates.

Also need to understand why customers don’t expand: Awareness, Value, Conversion.

  1. Awareness: look at whether they’ve seen the pages where they would learn about expansion paths, send surveys to see if they are aware, live screen, etc.
    • Analytics and instrumentation is helpful here.
  2. Value: they’ve seen the path (aware), but don’t take the action to expand.
  3. Conversion: friction to convert – try to expand but end up giving up or can’t.

Goal to expand revenue via expansion paths

  • Deepen revenue from current use cases (more frequency, more value per order, etc.)
    • Easiest, they are aware of the value.
    • Examples: Uber increasing ride frequency, Amazon getting you to buy more products e.g., with “buy it again” or “add all 3 frequently bought together to cart”. Adds social proof and lowers friction.
    • You may still need to invest in further activation. For example, if your expansion path is to add new users to the plan, each new user needs to successfully activate to drive the value and not risk churn.
  • Move customer to a higher ARPC use case (uptier, etc.)
    • Requires educating, converting, and activating the customer on the new use case.
    • Ensure that you educate (build awareness) in product, then target the right users based on existing information.
    • Lower friction to upgrade. For example, in Slack any user can trigger an email to the admin asking them to upgrade.
  • Add on more use cases
    • Requires educating, converting, and activating the customer on the new use case.

Get on more usage, a new use cases or building on new add-ons

At-risk customers

  1. Identify them by looking at:
    • Negative actions: clear actions that signal problems. Detractor NPS scores, customer support activity, visit to cancellation pages, features like exporting all data.
    • Negative Direction of engagement: is engagement decreasing?
    • Prediction model
  2. Understand why at risk
    • Satisfied but no longer need the product
      • What to do? Move them to a new use case.
    • Unsatisfied – seeking an alternative
      • Signals: Big drop in key actions (found alternative), left negative feedback, log customer support cases
    • Losing habit – external factors lead them to use less
      • Signals: Gradual drop in usage (losing habit)
  3. What can you do?
    • Habit reinforcement (e.g., Duolingo – manufactured triggers reminding you to practice and using rewards.)
    • Anti-conversion flow: decrease the perceived value of churning, increase perceived price of churning, increase friction to churn.
      • e.g., ClassPass in cancellation flow reminds you of benefits and show you classes that you attended. Instead of canceling they allow you to take a break. They add a reactivation fee in case you cancel. For cancelation they add a “contact us” flow.
      • Audible also shows benefits when you try to cancel. Also show you things recommended for you (part of memberships), incentivizing you to stay and discover something new. Then there’s a “no thanks” language to decline that offer to stay. Then they ask you “why canceling” to force you to answer.
      • Optimize to cancel those on the fence, don’t anger customers by adding too much unnecessary friction.
    • Use case transition
      • Uber from UberX to Uber Pool, Bumble from Bumble Date to Bumble BFF.

Churned customers

  1. Define
    • Not using the product and/or not paying for the product. Time they last placed an order, Last booking with a host, turn off auto-renew, cancel monthly subscriptions.
    • Time frame
  2. Identify candidates for resurrection
    • Figure out the time frame that matters. To do this, define different time periods and look at how many customers have been inactive for that time period or longer in the past, and of those, how many never are currently dormant (didn’t return) vs. how many returned.
    • Pick a timeframe with a probability for resurrection when they have a chance to return (not too low like <5%). Use both qualitative and quantitative data.
  3. Why they churned? Look at actions, error logs, etc. or qualitative data (NPS, etc.) to know why they churned.
    • Satisfied
      • completed a use case
      • What do to?
        • Use case transition (e.g., Trello for wedding user can be told about Trello for work).
        • Consider timing and channels to get them on the new use case.
    • Unsatisfied
      • Discounts/competition
      • better experience in another app
      • annoyed at upsell/upcharges
      • What to do?
    • Involuntary
      • forgot to update cc details –> easier to resurrect
      • forgot about the product
      • What to do?
        • Mitigate by minimizing conversion friction.
  4. What to do?
    • Reactivate – build an easy path for the customer to get back into a good state in the product and establish the key habit. For example, in farmville when someone comes back after a long time, their farm is fully dead, so there’s a simple one-click option to restore/clean up the farm and start from a healthy place.

Optimization Equation

Optimize so that: Perceived value > Perceived price + Friction

  1. Perceived value:
    • For example: Someone with a friend referral has a higher perceived value of the product than someone who clicks on a facebook Ad. Someone who’s been using the product understands the value more.
    • To increase perceived value look at: trust/credibility/authority, social proof, urgency, scarcity, belonging, completion
  2. Perceived price:
    • Perception is influenced by alternatives (other products, other plans, non-discounted price, etc.)
  3. Friction

Stages of optimization

  1. Educate: learn the target state and how it will solve their problems, show them the value prop, help them understand the features.
    • Ways to educate:
      • Tell: In-product messaging, educational emails,
      • Show: Explainer videos, Product demo, case studies
      • Experience: Free trials, free credit promos (more resource intensive and cost – but necessary if low starting perceived value)
    • Figure out the message (what are you saying?), the channel for the message (in product, ads, people, etc.), and the timing of when to educate them.
    • Potential customers:
      • Fremium: conversion to paid will be based on the delta between free and paid features/capabilities.
      • Direct to pay: focus on full cost vs. all features.
      • In Spotify, if someone is on free, they offer a shorter paid trial than if they are direct to paid trial.
      • In ClassPass, someone from an Ad gets a better trial than someone referred by a friend (assume higher starting perceived value from referral)
  2. Convert: move them to paying customers
    • Conversion flow (number of steps, information collected), payment methods, currencies, error handling.
      • Reduce friction by showing a “more popular” plan
      • Reduce friction by showing logos of companies on each plan (social proof)
      • Clarify what types of team normally benefit from a plan.
      • Reduce conversion steps.
      • Support multiple payment methods
  3. Activate: help customer establish a habit that adopts the new features they paid for. Otherwise, we risk them churning after not realizing the value of their purchase.
    • To activate customers:
      • Get the setup moment right.
      • Get customers to experience the aha moment of that new experience/action.
      • Establish the habit.

Once a customer has been educated, converted and activated, and on a healthy state, we can focus on expansion paths. See “Healthy state customers” section above.


  1. Reduce model friction temporarily.
  2. Problematic:
    • Discounts lower value over time, e.g., J Crew frequent discounts will mean that customers now expect the discount for future purchases. Zynga too. Udemy runs a lot of 70%+ discounts, so now they are seen as low cost. If they stop, problem.
    • Growth loops: if you acquire customers via incentives, low retention.
    • Cost of revenue: think of margins.
  3. Use incentives to:
    • Convert potential customers to existing healthy customers
      • Thumbtack offering new pros free credits to contact 10 leads. This would help them establish that first conversion.
      • Netflix Free Trial, etc.
    • Expand revenue from existing healthy customers
      • Volume discounts (e.g., ClassPass credit purchase get cheaper the more you buy)
      • Loyalty programs
      • Longer term (e.g., pay for a year)
    • Save at-risk customers and make them healthy again
      • Repair streak loss by getting duolingo pro.
      • Anti conversions like classpass reactivation fee.
      • Incentive to transition to new use case
    • Resurrect churned customers back to a healthy state
      • incentives are good for use case transition, but not for mitigation.

Challenges of Monetization Strategies

  1. Data is incomplete without testing
    • qualitative and quant research doesn’t fully capture real customer behavior.
    • research only covers a part of the piece, but there are many moving parts.
    • analyzes have errors.
    • hard to predict impact on other growth metrics.
  2. Testing is hard:
    • Customer sensitivity
    • Cross population: untested customers can find out about the test price
    • Time: tests would need to run for a long time in order to get full results, considering renewal/churn/etc.
    • Infrastructure lift to test

Case Studies


  1. Focus on Revenue and Gross Merchandising Value – gives you a sense of the overall pie or opportunity AND what piece you’re generating.
  2. Segmented data by cohorts:
    • Channel: directly, e-comm platform like Shopify/Magento
    • Segment: casual sellers, etc.
    • Size of order value
    • Transactions by AOV (avg order value)


If you see people abusing your platform, or using it in a way that was not originally intended, pay attention to see if there’s a use case there to develop and monetize. (e.g., businesses wanted to engage).

To match the market, analyzed on frequency of use and $$ value.


  1. Don’t let executives drive you, have multiple ways to convince different people:
    • Numbers: financial
    • Quotes and User Research
  2. Look at what’s in front of you to analyze, but time box it.


  1. SMB/self-serve needed much cheaper alternatives.
  2. Added monthly payment options (not just yearly)
  3. Simplified their billing to be more predictable by changing the value metric (instead of per event they started billing per “trackable users” for these users)
  4. Improved the free product to show the value of upgrading more easily (e.g., show “lock” state for features that were not available in free)
  5. Key takeaways
    • Activation journey is the key to healthy monetization
    • Monetization awareness is where 80% of the opportunity lies
    • Launching self-service monetization in a sales-led company is hard.


Last Updated on May 3, 2023 by Omar Eduardo