There is an excellent story behind Snapchat. The app which is mainly known for its filters has many reasons to invest in it. We all know people who use Snapchat just for these filters to take good and fun looking photos. But is it just for these photos, why exactly is the app investing billions on these filters, and how does it actually work?
Let us answer all those questions first there are broadly three reasons as to why Snapchat has kept those filters:
- To increase user engagement
- Get competitive edge
- Revenue generation
But before we dive into these three points, let us understand —
How does the tech behind Snapchat filter work?
We will decode the process one step at a time rather than reading plain facts.
Imagine you are going to make a filter for Snapchat, which uses a mobile camera to apply these features.
The first step would be to understand how the feature is expected to work, i.e. you need to define your goal. Here the filter is supposed to put objects/illustrations virtually on users face, like having puppy ears and tongue sticking out of the mouth.
So the process would be — a user opens the camera, the camera would detect the face, and the filters would work their magic on the face. Now the first step would be to make sure that images captured by a camera can actually be understood and differentiated by the app.
Here we use Computer Vision and the exact area of computer vision used is called Image processing. This is the same tech which Facebook uses to tag your friends in a group photo. For a computer, an image is a collection of 0 and 1, below is an image of a person taking a selfie.
As you see, there is a fundamental problem as you are unable to detect nose or any facial features. So the image is broken into tiny sections and each section is given a value from 0 to 255, where 0 means black and 255 means white. This is done to understand the facial features common in every face like -forehead is lighter than the eyes, the nose bridge is lighter than the sides of the nose, etc.
And once you replace the binary numbers with these pixel values, this is how it looks.
After this, Viola–Jones algorithm is used to broadly distinguish a face and the objects present. The algorithm uses contrast between light and dark pixel areas to match and verify whether it matches with a human face or not.
If you are a dying tech fan there are two specific algorithms used to detect your face called — HOG(Histogram of Oriented Gradients) and SVM(Support Vector Machine), you can read more about it.
So after face detection, you need concrete structure of the face, to understand where to put the tongue, otherwise if you see a tongue emerging from above your lips and below your nose, that would look bad.
This is done by facial landmarks. It uses data from millions of faces to understand the structure and exact boundaries of your face, it does so by matching your face with an existing dataset and then plots points onto your face.
Now it is essential to understand that no two faces are the same and so points never exactly fit, instead the algorithm adjusts according to your face. Then comes the actual image processing part. It uses millions of photos and machine learning to create a virtual face.
This virtual face aligns with the actual face and creates a 3D model around your face, which moves as you move, and this is also the model which detects whether your mouth is open or closed.
This process is known as active shape modelling and plays an important role in applying filters. But remember, this complete process is performed in milliseconds, and so far Snapchat has invested billions in its AR technology, improving it every day.
Okay, enough with the tech! Let’s move to more fundamental questions.
Why does SnapChat keep these filters?
From the product point of view, it solves some good problems and results in positive metrics for the company such as:
- Increasing user engagement
Before we go to user engagement, have you noticed that Snapchat is rather difficult to understand for a new user, the design is termed as hard to use by many first time users. It goes against the notion of intuitive design which says — make an app as simple to understand as possible because if users start discovering features right from installing the app, it makes them comfortable with the app and results in more usage. Simple and straight forward.
However, Snapchat follows shareable design, here the features of the app are not so “intuitive”, and you mostly learn the app by watching your friends, who show and teach you how to use these features.
Now shareable design is also said to increase user engagement because you spend time figuring out the features and once your friend shares some features with you, you take them as a “widely acceptable” or “cool” feature. The only problem is, many new users drop-off(uninstall the app) because they say it is complicated to use/understand. A risk, which could either mean high user engagement or early drop-off.
The battle of intuitive design vs shareable design is debatable, while I prefer the intuitive design, what are your views?
But the main reason for user engagement in Snapchat, as most people would guess is — filters. Unsurprisingly women use Snapchat more than men, and the main reason to use these filters are to play and explore.
Also, the number of Snapchatters submitting new Lenses through Lens Studios is growing at a decent rate. The creators of the lens also make money when advertisers ask them to make a specific lens based on their ad campaigns. Snapchat wants an army of lens creator so that advertisers have a wide variety to choose from.
Because as an advertiser makes and runs these ads, it would mean money in the pockets of Snapchat.
2. USP and Competitive edge
Filters are the USP(unique selling point) of Snapchat after all Snapchat identifies itself as a camera company while describing in its company profile.
The app in general is designed such that users report feeling silly, creative, and flirtatious after using it. While after using Instagram, users feel entertained and flirtatious, but “self-conscious and insecure” at the same time.
Hence, the app wants a user to feel positive emotions after usage. This is also in line with its strategy to focus more on user engagement than gaining new users.
Yes, Snapchat understands that competing with Instagram for active users does not make sense, instead focusing on retention and user engagement is more profitable.
A good area of focus should be developing filters which are silly/entertaining, that involves kids participation aged between 5–9 years. The app is already easy to use, and parents do not generally have problems with kids using Snapchat. On the other hand, Facebook and Instagram can expose kids to adultery and wrong information, but Snapchat is mostly a feel-good app.
This would mean early onboarding of users, as once these users(kids) come of right age 18–30 years, Snapchat would be their default app for photos/entertainment.
Another good USP of Snapchat is the discover feature and again it is more focused on entertainment that adultery. While Instagram has a similar feature in terms of IGTV, but it still hasn’t got the traction, it could possibly have considering its active user base.
Another indicator of having good USPs is that Instagram keeps copying Snapchat, from Stories to direct messages to filters; it copies a lot. While copying does prove that its features are good but it definitely hurts its revenue.
3. Revenue generation
In 2019, 98% of Snapchat revenue came from ads, which isn’t surprising considering its a free platform that uses customer data to generate revenue.
There are a number of ways in which Snapchat generates its revenue, and advertisers pay crazy money to get those ads because Snapchat has access to a premium age of audience that none others have. The age between 18–30 years, leveraging this Snapchat has made several revenue streams, like:
- Snapchat uses its ‘Discover’ feature to run ads, and these cost around $50k per day. Yes, the amount is crazy and advertisers pay money for it. The price has been fallen from its previous highs of $750k per day. These ads are at premium places and are generally reserved for big brands.
- Snap Ads — these are mobile video ads with a choice to add interactive elements to the ad, such that you can make a game within the ad, and the click-through rate for these ads is much higher than the industry average.
- Snap also allows six-second unskippable video ads which appear during Snapchat Originals episodes, or during Snap Games.
- Snap is also investing on third-party integrations via Snap Kit which allows users on apps like Tinder, Netflix, etc. to create custom partner stories from Snapchat. Partner apps can then run Snapchat stories on their app, and could even monetize content with full mobile ads.
- And last Sponsored lenses, one of the most expensive offerings in Snapchat, the rates for which range from $400k up to $700k on special occasions such as Super Bowl. One of the most talked lens, a Taco Bell’s lens was viewed 224 million times. That’s huge.
And with this we come to an end, some things worth noting is that Snapchat has a high user engagement and ads targeting, which allows them to charge such high rates for advertisement.
The move ahead is to focus more on user engagement because competing with Instagram on number of active users is quite difficult.
These are my and Nasir Afroze’s view on the topic. Do let us know what you think.
