Don't miss out
Subscribe to get the latest from Accel emailed directly to you
No framework is one-size-fits-all. We need a playbook which anyone can tailor for their businesses...
No framework is one-size-fits-all. We need a playbook which anyone can tailor for their businesses. In this post, we will see how you can monitor the process of building economic moats/value creation mechanisms with the help of metrics for your company. This post is a part of the Growth Spiral series, read more about it here.
If you’re not convinced about why you should use this framework, read the previous post on where the current frameworks fall short –
Why you need the right metrics for your startup
Firstly understand the problem category your business is in. Each problem category will have multiple actions associated with it. For example, in the ‘travel’ problem category, there can be many actions like ‘taking a ride’, ‘renting a car’, ‘buying a car’, ‘servicing your car’ and so on.
The action undertaken by the user for which they primarily come to your offering is the ‘core action’. In the travel problem category, for a ride-hailing businesses, ‘taking a ride’ becomes the core action. However, for a car-selling business, ‘buying a car’ becomes the core action. From the core action, we can estimate the maximum frequency of usage we can expect from our audience, which we call ‘natural frequency’.
Natural frequency is the frequency at which the activity happens organically.
The next step is to define the segments of users across different phases of the user journey based on their definitions as explained in the below examples.
Let’s take the example of a food business. For the sake of the example, let’s consider this is as a business which lets customers order food from its software application. The core action is ordering food and the natural frequency is 3 times a day.
Let’s take another example of a task management software. The core action is creating tasks and the natural frequency depends on the kind of customers you have. For the sake of the example, let’s consider this software is primarily used by small to medium sized businesses which would have at least 5 tasks a day, which is 25 tasks a week.
Let’s take another example of a business which has a low natural frequency of usage. Consider a healthcare provider aggregating software helping patients consult doctors efficiently. The core action here is the patient consulting with a doctor. It is said that an average person goes to the doctor at least once in 3 months, which means the natural frequency is 4 times a year. We will use alternate terminology for user segments which would be more relevant for a business solving ‘important problems’ (as opposed to ‘frequent problems’).
Note: User segments can be defined for each action separately. For example, the same hospital-aggregator company might also have another business line where they are helping doctors find patients efficiently — that business will have it’s own user segment definitions and hence a different growth score eventually.
Like in any process optimisation, we need to define the ‘how much’, ‘how good’ and ‘how fast’ metrics basis our first 2 steps. The reason we have 2 quality metrics for each phase is to measure the 2 folds of ‘how good’ — ‘how frequently’ (to factor in natural frequency) and ‘for how long’.
Stickiness score is a quality metric which tells us ‘how frequently’ are users using our product.
The absolute benchmark for the measure of ‘how frequently’ is natural frequency.
It has been discussed multiple times in the past by Social Capital in the form of L28 Distribution and also by Andrew Chen in 2018 in the form of The Power User Curve. Here is the formula for stickiness score –
Stickiness score = [(1 * % of users who are active for 1 day of the month) + (2 * % of users who are active for 2 days of the month) + … (30 * % of users who are active for all 30 days of the month)]/30
Quick ratio is a quality metric which helps us answer the question — ‘for how long’. We look at the quick ratio of acquisition, engagement and expansion to understand the quality of users in each segment is. The better the quality of conversion is, the longer the users will use our offering.
Quick Ratio = (New + Resurrected)/(Lost + Contracted)
For example, if quick ratio of acquisition is good, that means much more new users are activated as opposed to the existing activated users stopping usage of our offering.
Applying this to the 3 phases, we get the following –
We typically measure these metrics on a monthly basis, but measuring them on a weekly basis will only give faster feedback. In the example we’ve taken in step 2, we just replace the definitions of the user segments to the respective phases. Now that we have defined the base metrics, let’s measure aggregate numbers to make the framework simpler and easier to use.
Note: In some companies solving ‘important problems’ (especially in one time purchase problem categories), the frequency of choice (or purchase) is less, but there will be other actions in the problem category which indicate the usage of the chosen product. While calculating the quantity, quality and velocity metrics, it’s important to consider the frequency of usage and not choice since both will be correlated, however the former will be a leading indicator to the latter.
Recall that businesses solving ‘important problems’ would focus on better brand connect (or ‘Connect’), as opposed to those solving ‘frequent problems’ which focus on better scale (or ‘Invite’). To have a measure for each of the 3 phases in the customer journey separately would help us understand how our business is doing relatively across these phases.
Invite score = (% change in quantity metrics + % change in quality metrics + % change in velocity metrics)/3
Percentage (%) change can be measured across any time period — month-on-month, year-on-year, etc. Similarly, we can derive Engage and Connect scores as well.
The natural frequency has been factored into the metrics we defined in step 3, this enables us to understand how to maximise the relevance of our business offering to the target audience. The next step helps us in understanding how the success of the initiatives undertaken are reflected in the subsequent cohorts and hence capturing spirality of growth. We measure growth score as an improvement metric which would tell us holistically how well the company is growing.
Growth score = (Invite score + Engage score + Connect score)/3
Typically, when the natural frequency is high, we unlock and capitalise on scale as a moat — which will show in a higher ‘Invite score’. When the natural frequency is low and the importance of the activity is high, we unlock and capitalise on brand connect as a moat — which will show in a higher ‘Connect score’. You can surely look at weighting factors for each of the 3 scores based on whether your company is solving ‘frequent problems’ or ‘important problems’. We will publish benchmark weights subsequently when we are analyse data across a larger number of companies.
Let’s consider the example of a food business which lets customers order food from its software application. Say an experiment of incentivising people to review restaurants was successful and you designed the review system into your product. Because of the reviews, existing users start ordering more confidently and have a tendency to get closer to the natural frequency of ordering food. This results in a spiral loop reflecting in the engage score for the next month.
Let’s consider the example of a business selling cars through an online listing. Even though the purchase of the car is the core action of choice, we will consider the core action of usage — using the car/taking a ride. Say an experiment of encouraging customers to refer other people to try your application was successful and you designed the referral system into your product. If the referral system is successful, super customers will invite high potential leads. This results in a spiral loop reflecting in the invite score for the next month.
Looking at the Growth score on a monthly basis will give us the high level indicator of whether we are progressing as we expected or otherwise. If we realise that course correction is needed, we can double click to check which phase is where we are falling behind and go even deeper on figuring out whether the quantity, quality or velocity is taking a hit.
Moreover, as the community derives more and more value, we will have a central repository with benchmarks across the metrics to refer back to. If you’ve already applied this framework to your business, or need help applying, I would be excited to get in touch with you at firstname.lastname@example.org.