I hope you and your family are doing well!
It has been some time since the last post. The reason for that is something exciting I am going to reveal in the next 2-3 weeks :)
Before we start, I have an update to share. I am going to do an online session around Product-Led Growth this coming Wednesday. Given that you are reading Growth Catalyst, I thought you might be interested in this.
The event has seen a good response and has gotten 1,000+ registrations so far :) Here is the link to register for the event.
Off to the topic at hand,
Strategy in the Internet era is different from most of the 20th Century. The fundamental advantages of the Internet are three-fold:
Network effects (discussed in the last post)
Data generation and computation power
The speed at which the information travels, providing real-time connectivity and communication
The second-degree effects of these fundamental advantages are varied and powerful.
Let’s take Uber, for example. Uber is a marketplace where riders can find drivers at a particular cost. The network effect of Uber is shown in the image below.
Apart from the network effect, Uber also uses the data for past rides in multiple ways:
Rating data to remove bad riders and drivers to ensure the quality of the marketplace is good.
Calculating ETA based on past travel time on the same route. A more accurate ETA means a better experience for the riders.
For any new company starting today, it will take some time to collect data and create an experience as good as Uber. That creates a barrier to entry.
The last advantage that we talked about is the speed of information travel. Because the data from all the rider and driver apps can travel instantaneously (< 1 second), the data can be used to do real-time matching of drivers and riders. The real-time demand data can also be used to calculate surge pricing. We can also use real-time location tracking to build safety features for both riders and drivers.
All of the above makes Uber a far superior experience as compared to hailing taxis offline. And that’s the reason for the > $100 B market cap of Uber today.
Now that we have understood how the fundamental advantages work for an Internet business like Uber, let’s understand more about what common strategic advantages we can build over traditional as well as other Internet companies. We have already discussed network effects, so we will focus on data and real-time interconnectivity in this post.
Data is the New Oil
Data refers to pieces of information. The data can be a photo, a piece of text, information about a person like their name and phone number, etc. We use data and information interchangeably in everyday life.
Internet companies like Google and Facebook generate lots of data from us. For example, Facebook knows our name, how we look, our friends, etc., through our Facebook account. Google knows what we are searching for and which websites we are visiting. As these companies learn more and more about their users, they can use the data for advertising the products that users may buy. Say, a user searched "how to complete a marathon?" on google.com. Google can show the user advertisements around running t-shirts, bottles, shoes, etc.
Based on their users' data, Google and Facebook have built the largest advertising platforms globally. From the ad, Google and Facebook make hundreds of billions of dollars in revenue every year.
But just like crude oil, data in itself isn’t useful. We need to process and transform this data to put it to use. The good news is that the storage and computation costs of data have come down. Since it’s cheaper to store data and process it, not doing it well can be a big disadvantage.
Unlike oil, data is an infinite resource. We keep generating data as long as we live. The companies that can record and analyze this data to create products rule the world.
Usually, data and real-time connectivity work hand-in-hand to build advantages.
Because the data/information can travel within 100s of milliseconds on the Internet, it can connect two devices sitting across continents in real-time. This could lead to
information exchange between company servers and user devices seamless
information exchange between two users seamless on social media or chat applications
aggregation of data in real-time to build valuable experiences like community, live events, etc.
real-time feedback and correction in a process
The list goes on… When coupled with data generation and processing, real-time connectivity can add a great advantage.
So how do you systematically think about using data + real-time connectivity for your own app/website? There are four important aspects to think about:
Data generation activities
Feedback loops from consumers
Machine learning to automate decisions at scale
Connecting various players to create better coordination
1. Data Generation Activities
Data generation activities are the first place to look at. Most companies miss multiple data points in the value chain and experience due to either lack of thinking or heavy investment needed. The ones that invest in it reap the benefits heavily.
A company that is amazing at it is Shein. It is an eCommerce company out of China that is present in every major market in the world. It reportedly did almost $10 billion in revenue in 2020 and has grown over 100% for each of the past eight years. Shein did this by cracking fast fashion. Fast fashion was pioneered by Zara in the 1990s when it started churning out affordable fashion clothing based on the newest fashion trends within weeks. Shein took it to a different level by making it almost real-time.
Shein generates fashion-related data from multiple sources like google trends, competitor sites, etc. But what it did quite differently is Shein’s supply chain management (SCM) software. Factories that work with Shien have to run on its SCM software.
Garment factories are not typically digitalized. By digitalizing what was happening on the factory floor, Shien started generated the data which no one had earlier access to. This would mean that its SKUs on the website can be updated in real-time and much more accurately. This also means that factories get better demand forecasting as well. This is how Shien’s retail flywheel looks like (source: Matthew Brennen),
GPS systems installed in cabs and bikes to generate real-time tracking data is another example of data generation.
Figure out where you can generate data that you are missing. Because of real-time connectivity, businesses that can connect different parts digitally and generate data can succeed more than those that can’t.
2. Feedback Loops from Consumers
Steve Jobs famously said, “You've got to start with the customer experience and work backward to the technology.”
Feedback loops can be built into the product to make it better. One such example of a feedback loop is ratings and reviews. Ratings work as social proof for consumers, and they are more prone to buy something that other people have already bought.
You also have to design the interfaces in a manner conducive to generate data. Facebook initially had just user-profiles and no newsfeed. Facebook feed created a way for users to like, comment, and click on a post, which created engagement and revenue opportunities.
Auto-suggest wasn’t available in earlier search engines. Google auto-suggest feature takes into account whether the user clicks on top suggestions. If the user didn’t click on any of the top suggestions, that goes back as feedback to the auto-suggest algorithms.
Recommended product click-and-buy is another feedback mechanism for eCommerce websites and ad products.
Fitness apps of the future will take into account users’ sleep quality, blood glucose levels, heartbeat, and energy levels as the feedback to recommend the right amount of workout.
All of these examples go on to show that you can create products in a way that is more conducive to receive feedback from users.
3. Machine Learning at Scale
ML/AI automates business logic at scale. Want to figure out what to show to customers that they will like best? ML does the job by creating a recommendation system
Want to figure out which of the leads are most susceptible to convert into a paying customer? ML does the job by assigning a lead quality score. Want to figure out which of the customer are going to churn? ML does the job.
You get the picture. The problem with ML is that you can’t apply it to all the problems. You need to have lots of data to solve a problem through ML. I would devote a future post around how and when to use ML to solve a problem.
4. Better Coordination
What’s the biggest difference between a Shein and a traditional fashion store? It is the real-time coordination between various players in the supply chain - store owners, factories, logistics/transport, etc. Real-time coordination helps players correct their course quicker and avoid waste of time, money, and resources. Over time, a more efficient supply chain can create cost leadership - one of the key strategic advantages.
Real-time interaction between online events and audience on Zoom, Facebook Live, and Youtube lets the creators adapt according to the audience response.
The same can be applied to offline sellers in India who can now send images and videos of the product to their long-time customers and get feedback on their inventory. In this way, they can avoid lots of inventory that isn’t in demand. Further, they can forward this feedback to their suppliers who can change their investments on certain SKUs. It’s beneficial for everyone in the network and is possible because of the Internet.
The businesses that use the key advantages the Internet has to offer will go on to create a stronger moat and competitive advantage over those that don’t. Over time, the faster cycle of data, feedback, and coordination results in faster evolution of the product. Internet products today are changing at a faster rate than ever because of these reasons.
So what changes are you going to make in your product and business, now that you know how to build competitive advantages?
This would be all for the post. See you next time,