Using Metrics To Drive Product Growth

Read the headline again. Isn’t it silly?

You are now probably thinking — How could we even think on product growth initiatives without using metrics? Silly guy.

I absolutely agree, so what’s the deal? BTW, I’m not concerned on sounding silly, smarthead!

Yes, everybody seems to be a data nerd nowadays. There are three types of them: the specialists, the trend followers, and the ones who ask questions. However, what I noticed is that the first two are the largest majority today, which makes the conversation around metrics be too technical. In this context, product teams tend to get stuck into a loop of data complexity and forget to zoom out and ask themselves — What metrics are for anyway?

The purpose of a team is to have impact on growth, which means get and keep more customers. Using metrics is fundamental for a product to achieve Traction and it is in the centre of the team’s two major goals:

• Driving value to users — building the right thing faster and cheaper.
• Turning users into passionate customers — identifying the most effective ways to build and engage the user base using a process of rapid experimentation across marketing channels and product development.

In that way, data is used to measure business performance, to analyse cause-effect experiments, and to build a habit-forming product with behavioural segmentation. It’s the glue between business intelligence, user tracking analysis and customer relationship management tools.

The way a team use metrics also changes according to the stage of the product lifecycle. For example, if you are in the Proof-of-Concept phase, you will concentrate the majority of your attention on qualitative data; however, if you are in the Traction to Scale stage using quantitative data is key.

What makes the difference in being a problem-solving team is to have the ability of sensing uncertainty and understanding which questions we need to answer before diving into building. The more you know about your users, product, and channels the better you become at coming up with solutions that influence growth.

To impact growth a team must be outstanding on using quantitative and qualitative data to identify the biggest problem or area of opportunity and apply it to prioritise its activities smartly. Knowing which metrics to use for each particular situation and how it can actionably be used to drive growth is absolutely key.

In a very simple way, using metrics to drive product growth is based on analysing and segmenting the user behaviour through the Customer Lifecycle and the Customer Profile.

Customer Lifecycle (source: http://leanstack.com/)

Customer Profile

First, we focus on identifying what kind of customer performs best and gets more of them. Next, we figure out how to move users from one stage to the next by identifying what factors turn users into active ones and optimise it with A/B Testing. Finally, we concentrate our efforts on building a habit-forming product that works deeply in the user behaviour.

To exemplify the application of these guidelines in a real situation I am going to use the following hypothetical scenario: the product is a graphic-design SaaS tool with a freemium price model and the team must double revenue in the next 12 months.

The following content describes how an experienced team will use metrics to achieve that goal. The initiatives and examples are structured into four stages:

1. Establishing a baseline
2. Focusing on the customers who perform best
3. Identifying what factors turn users into active ones
4. Building a habit-forming product

Initiative 1 — Establishing a baseline

I recommend an approach that starts by establishing a baseline of metrics, which is the starting point for developing growth optimisation. We must first understand where we are now and where we want to be, and then think on the most effective approach to achieve the goal and to be able to prioritise smartly.

Actions:

• Establish key metrics and actionable reports
• Basic segmentation

Key metrics and actionable reports

KPIs:

• MRR (Monthly Recurring Revenue)
• ARPU (Average Revenue Per User)
• LTV (Lifetime Value) — an estimation of the aggregate gross margin contribution of the average customer over the life of the customer
• Paying Users
• Churn — the percentage rate at which customers cancel their recurring revenue subscriptions
• CAC (Customer Acquisition Cost)
• Months to Recover CAC

Source: https://blog.kissmetrics.com/

Customer Lifecycle Performance:

• Acquisition
• Activation
• Retention
• Referral
• Revenue

Activation Performance:

• Conversion Funnel
• Conversion Rate Cohort

Retention Performance:

• Repeat Cohort
• User Base Overview
• NPS — Net Promoter Score

Weekly Cohorts:

• Weekly group of users (by acquisition date) compared by the Customer Lifecycle Performance

Basic Segmentation

Segmentation options (applied to all the metrics/reports above):

• Traffic source / medium (email, social, paid, organic, direct, referral, etc.)
• Demographics
• User type (new vs. returning, mobile vs. desktop, frequent vs. infrequent, long vs. short sessions, multiple sessions vs. single session)
• Features used / content viewed

Tools Examples

• Amplitude, KISSmetrics, MixPanel
• Segment.io
• Tableau
• Google Analytics

Initiative 2 — Focusing on the customers who perform best

Targeted Acquisition Campaigns

Once we understand which segment of users are performing best in our Key Metrics, and where they are coming from, the fastest way to grow is to focus on acquiring more of them, using the channels that are performing best.

Referral Campaigns

Referral Campaigns are a very effective way to “acquire more of them” and have the potential to influence growth exponentially.

Targeting the Customer Profile:

• Target the kind of user who is performing best
• Target users by their feature of preference or specific content of interest
• Segmentation options: topics of interest, content type, age, localisation, profession; time, week day, important dates; devices — web, mobile, tablet; source — email, social, paid, organic, direct, referral, google, facebook, twitter, etc.

Targeting the Customer Lifecycle (actions triggered by the Journey Stage):

• Target Active “trial users”
• Target Active “paid users” (basic)
• Target Active “paid users” (pro)
• Target Active “paid users” (enterprise)
• Segmentation options: Journey Stage — unaware, aware, interested, first-time customer, regular customer, passionate customer; Frequency (level of activity); Customer Level (free, trial, paid).

Possible incentives to trigger the user action — referring the product to a friend:

• Trial extensions
• Unlock paid feature for limited time
• Additional storage for free
• Access to special templates
• Gift Card to buy photos, illustrations, and templates
• Exclusive educational content/course
• Exclusive promotional item
• Large discount: 10% discount for friends / 20% for me (the user)

Optimising the Basic Reproduction Number (or Viral Coefficient):

A/B testing experiments to validate cause-effect factors that improve the Basic Reproduction Number of a specific Referral Campaign.

Sales Representatives

Free Trial leads:

• Lead Management: trigger a notification to a Sale Representative (using a tool like Autopilot HQ, for example) when the Free Trial deadline of an user is coming closer.

Same Customer Segments and Channels, targeted Approach:

Keep the focus on targeting the kind of user who is currently performing best (for example, Online Marketers) and on using the same channels from where they are coming from.

• Optimise Channels (inbound and outbound) and Communication Touch Points to the particularities of small business users in the Online Marketing industry.
• Offer special benefits for small business users in the Online Marketing industry (for example, “Prezi” and “Evernote” tactic on offering discounts to Students when signing up with a school email address)
• Lead Management: once a lead with this target characteristics is converted, a Sale Representative specialised in the Online Marketing industry receives a notification in Slack (using a tool like Autopilot HQ for example) and reaches out the potential customer.

Same Customer Segments, new Channels:

Keep the focus on targeting the kind of user who is currently performing best (for example, Online Marketers)

Business development to boost targeted lead generation:

• Establish partnerships with other SaaS tools in the market which also target small businesses in the Online Marketing industry
• Establish partnerships with Online Marketing community influencers
• Sponsor Online Marketing events which target small business professionals

International Markets Acquisition

Global expansion is not a small effort, so prioritisation is key. In the Initiative 2 we had identified what kind of user is currently performing best, where they live, and focus on getting more of them. The idea now is to use the same tactic, but narrow the focus per country/language.

• Which are the countries/language that are currently performing best?
• What kind of user is currently performing best, where they live, and focus on getting more of them (in each particular country/language)

Which are the countries that have the potential to perform best?

• Market attractiveness: markets with high broadband and mobile penetration, favourable socioeconomics, stable political environments, accessible payment infrastructure and relatively easy regulatory and tax requirements.

The obvious action to take next is to optimise the Channels (inbound and outbound) and Communication Touch Points to the particularities of the different languages and consumer behaviour in each country.

Initiative 3 — Identifying what factors turn users into active ones

After having identified what kind of user perform best (using the metrics explained in the Initiative 1), we can analyse what factors turn them into Active ones, and use our learnings to keep current successful factors and optimising unsuccessful ones (consequently improving churn).

Optimisation Actions

Quantitative Data:

• Funnel Report: compare segments (best vs. worst), identify bottlenecks
• Repeat Cohort: compare weekly group of users by event frequency (daily, weekly, monthly retention) and analyse possible cause-effect factors
• Weekly Cohort: compare weekly group of users by Customer Lifecycle Performance and analyse possible cause-effect factors
• Compare weekly group of users by D1, D7, D30 user retention and analyse possible cause-effect factors
• Compare Product Features user retention/frequency using Cohort Reports and analyse possible cause-effect factors
• A/B Testing results

Qualitative Data:

• User activity tracking using People Search report, Heat map, Scroll map: user activity analysis of specific users that represent different personas (Best vs. Worst)
• Customer Journey analysis of specific users that represent different personas (Best vs. Worst)
• Interviews with specific users that represent different personas (Best vs. Worst)
• Analysis of the three behaviour elements of the Fogg Behavior Model (Motivation, Ability, and Trigger) with specific users that represent different personas (Best vs. Worst)

Experimentation:

• Use A/B testing experiments to validate cause-effect factors that turn user into Active Ones and optimise the product and communication

Pricing Experimentation

Quantitative data:

• We need to have Activation and Retention tracking metrics in place (vide Initiative 1), and A/B testing results.

Qualitative data (in-app survey, in-person interviews):

• Focus on understanding user reasons to pay (or not pay, downgrade)
• Understand their perception on value — Jobs-To-Be-Done, what costs The Product avoids vs. how much The Product cost (money, time, mental energy) comparing to other alternatives.
• Understand user’s budget priorities — sometimes we are competing with tools on other categories for the same budget share.
• Understand their perception on current price vs. The Product strategic positioning

Experimentation:

• Use A/B testing experiments to validate with quantitative data what we had discovered with qualitative data

Initiative 4 — Building a habit-forming product

Using Segmentation to improve Conversion, Churn and product subscriptions.

Once we had identified what kind of user perform best (Initiative 1), and what factors turn them into Active ones (Initiative 3), we can focus on building a habit-forming product experience that shapes user’s’ preferences and attitudes.

Goal:

1. Getting our users to use more our product
2. Increasing user engagement up to the point they understand the value to subscribe/pay
3. Getting our users to subscribe for our product
4. Ensuring our users continue to use and pay
5. Getting our users to refer our product

Building an engagement loop using the Hook Model:

1. Trigger: External (push notification, email, retargeting, etc) or Internal (emotions associated in the user mind)
2. Action: the simplest behaviour in anticipation of a reward. Three elements must converge at the same moment for a behavior to occur: Motivation, Ability, and Trigger (vide Fogg Behavior Model).
3. Reward: types of variable rewards: Tribe (search for social rewards — recognition, cooperation, competition), Hunt (search for resource — food, money, information), Self (search for self-achievement — levelling up reflect mastery and competency, task management reflects consistency and completion).
4. Investment: users “invest” for future benefits. Investments increase the likelihood of the next pass through the Hook loop in two ways: loading the next trigger of the engagement loop (example: installing a bookmarklet, downloading an app); Storing Value, improving the product with use — how more you use, more it get better (example: the more content, data, followers I have, better the experience).

Designing the Engagement Loop

• Trigger: Customer Lifecycle segmentation defines the Trigger (the user lifecycle stage defines when and what kind of trigger will occur)
• Action: Customer Profile segmentation influence Motivation and Ability that will trigger Action
• Reward: Customer Lifecycle and Customer Profile segmentation define the type of Reward that will best influence the user to move to the next stage
• Investment: Customer Lifecycle segmentation define the right investment to offer to the user

Customer Lifecycle Segmentation: Journey Stage — unaware, aware, interested, first-time customer, regular customer, passionate customer; Frequency (level of activity); Customer Level (free, trial, paid).

Customer Profile Segmentation: topics of interest, content type, age, localisation, profession; time, week day, important dates; devices — web, mobile, tablet; source — email, social, paid, organic, direct, referral, google, facebook, twitter, etc.

How to implement it:

1. Establishing Key Metrics using behavioural and transactional data (vide Initiative 1)
2. Segmenting users by Customer Lifecycle and Customer Profile: identifying what kind of user perform best (vide Initiative 1), and what factors turn them into Active ones (vide Initiative 3); selecting users sample for the validation/testing phase.
3. Designing the Engagement Loop
4. Validating Engagement Loop assumptions with the users sample (small scale): Quantitative: Key Metrics (vide Initiative 1), A/B Testing; Qualitative: in-app survey, in-person interviews, NPS.
5. Implementing automation triggers to all the user base: CRM (e.g. Amplitude + Autopilot HQ)

P.S.: implementing initiatives based on the fundamentals described above using Machine Learning is a way to boost the habit-forming product experience even more.

Conclusion

According to the author Warren Berger — A More Beautiful Question: The Power of Inquiry to Spark Breakthrough Ideas, “one of the many interesting and appealing things about questioning is that it often has an inverse relationship to expertise”.

Author Stuart Firestein, in his book Ignorance: How It Drives Science, argues that one of the keys to scientific discovery is the willingness of scientists to embrace ignorance — and to use questions as a means of navigating through it to new discoveries.

Keeping metrics simple and having an actionable recipe as a starting point proved to be an effective approach. However, what really makes a team have high performance in the long term is the ability to frame the work as a learning problem, recognise uncertainty, acknowledge your own fallibility, and model curiosity by asking a lot of questions.