I was shy of 20 years old when I started my company. I set up a cloud kitchen, a fancy term for a kitchen that receives orders, usually through the internet, prepares the dishes and then delivers them to their customer’s address.
My kitchen prepared North Indian health food. An oxymoron of sorts. Think of traditional Indian dishes but with re-engineered recipes to make them healthier. My target customer was a member of the middle or high-income segment looking for healthier food alternatives.
When we started, my company (Orange is the New Green) would receive orders, prepare meals, deliver them and do all the marketing/branding work in-house. There were three distinct activities that we performed—
a) Prepare food — Operations
b) Deliver the food — Logistics
c) Marketing and Branding — Growth
Growth meant scaling all three verticals. While there existed repeatable processes that would help scale the Logistics and Growth verticals, it was a bigger challenge to figure out how to scale up operations.
We had grown to three kitchens and it was already a challenge to ensure quality at this scale. There are two reasons for this, firstly, the food we made was complicated to prepare and difficult -though not impossible- to break up into repeatable, and consequently optimisable, processes. Secondly, the labour market was disorganised and it was complicated to set up an incentive structure that would encourage kitchen staff to ensure a high-quality dish each time — again not impossible, but very difficult to implement and scale.
This is when we pivoted the business model. Playing on my strengths, I realised that I had a valuable customer base and a strong last-mile delivery network. I decided to outsource the operations vertical to third parties.
We would co-invest with aspirational food entrepreneurs who would in turn sell their products to customers through a multi-sided marketplace product. I bet on product-led growth.
Creating and growing a platform is a very different business than the one I was experienced in. It required an entirely different perspective and thinking pattern. To start, I needed to deeply understand my purpose
(or value add) to the parties in the marketplace.
Using first principles thinking, I first looked at the marketplace model. Any marketplace exists to reduce friction. The friction involved in transacting with other parties in the absence of a marketplace. The quantum of friction reduced is directly correlated with the value created by that marketplace. It became clear that I existed to reduce friction between people who wanted to eat healthy food and the people who wanted to sell healthy food. Great!
The second order of business was then to define success. If I didn't know what success looked like, I wouldn't know what the company was moving towards. Which led me to think about the North star for the company and then the product.
The idea then became to identify a metric that would quantify the value created by my product. It was simple, an efficient marketplace facilitates more transactions than an inefficient one, following which, the number of transactions would reflect the value created by a marketplace.
This is what I decided would be our north-star metric and our guiding philosophy started looking something like this —
Tier II metrics: Frequency
The total transactions metric is easy to measure but doesn't yield actionable insights. For example, my customer orders went down in January 2021, so what? If I wanted a headline metric to unite the stakeholders of my company, this wasn't it, so I broke it down into Frequency and Quality. This article focuses on frequency metrics. The next one will deal with Quality metrics.
Focusing on Frequency shifts attention to customers’ decision matrix and distribution, which could potentially yield more insights. To understand transaction patterns, we segmented customers into cohorts depending on how frequently they ordered in a month. See below a sample customer distribution between March and Jun 2015.
Customers that fall into the >1 transaction cohort are those that transacted more than once in the period or repeat customers, the other cohort transacted only once.
This simple analysis can yield powerful insights. At the very least, it can lead you in the right direction. You can form hypotheses to test out for example.
In June, we launched a new “Subscribe” product feature that allowed customers to pay a discounted price upfront for the following 3–7 meals depending on the package they chose.
All things being equal, an increase in customers who order with greater frequency would indicate the success of this feature. This hypothesis can easily be verified by cross-referencing this data with users who opted for the “Subscribe” product.
Tier III Metrics: User Segmentation
If the goal is to create a human-centred product, it helps to understand the user behind the data.
It helps to double-click into a cohort to understand behaviour better. If we were to zoom into the >1 transaction cohort for June vs March, we get the following chart —
We see that as a result of the “subscription” feature, our >1 cohort doubled. More interestingly, we see that the shift of users has been from the “1 per week cohort” and the <1 per week cohort is unaffected.
This is actionable intelligence for decision-making. Action that can potentially tie together different verticals of the company.
My hypothesis was that the “1 per week” cohort is a feeder for the “>1 per week cohort”, a target cohort that I was seeking to expand. My priority would be to push users from the “Single transaction” and “<1 per week” cohort to the “1 per week” cohort. I will prioritise these actions since the new product feature is already effective in pulling in users to the “>1 per week” cohort.
With this direction, we can decide on a target for the Month/Quarter. Let’s say that we aim to increase the “1 per week” cohort by 25% in the next three months — by the month of September — I will then cascade this target to the operations, product and marketing teams to create an individual strategy and execution plan.
I will close out this article here. In part II of this piece, I will look at the various tiers of Quality metrics and the actions they inspire. In future articles, we will talk about how product metrics can help design operations, logistics or marketing strategies and execution plans. Stay tuned!
As always, would love your thoughts and comments. If you like my work, I would be honoured if you share it in your network :)
My socials are —
Linkedin — Link
Twitter — Link
