How AI Powered the FAANG Vendor Movement

Eric Repec Written by Eric Repec
What is the FANG Vendor Movement?

FANG is the acronym for the four most popular technology stocks in the market which have generated spectacular investor returns. These stocks are Facebook, Amazon, Netflix and Google. The term was originated by CNBC's Mad Money host Jim Cramer in 2013. Since 2017, Apple was added, thus it is now called FAANG. One can understand why they were called out as an elite group: 

They have a disproportionate influence on the overall valuation of the S&P 500 index and drive the the momentum for the overall market. For example, Amazon became a trillion dollar company upon dominating in E-commerce and cloud computing. 

A Hotbed of Opportunity

FAANG is synonymous with digital transformation, and AI is a key enabler for these technology leaders. For starters, FAANG vendors have tremendous footprints (e.g., Amazon has 150 million active customers; Facebook has 30% of the world's population as customers), and thus a massive store of data just waiting to be "learned" and put to good use. This data leads to opportunity-rich use cases that have the potential to drive game changing results. The FAANG vendor movement has influenced AI because of the tremendous business opportunities presented by these use cases and the rich data store required to hyper-focus in on the customer and make revolutionary improvements in how companies interact and service each individual customer interaction.

AI at the Heart of FAANG Greatness: Netflix and Facebook

Netflix exponentially grew its user base by nearly 7x, from 22m in 2011 to 150m in 2019. In 2015, they published a paper about how they transformed the customer experience with a personalization and recommendation engine, saving $1b in the process. The issue AI addressed is that a Netflix member loses interest after 60-90 seconds of looking for a title; when the customer abandons the service, their attention goes to something else and Netflix loses the engagement (which leads to cancellation of the service if it happens too often).  First, Netflix created a new KPI called "Take Rate" and identified a way to quantify it with Big Data.  Then they learned that the take rate was 3-4x higher when personalized videos were provided in under 90 seconds, versus just providing a list of the top watched videos. This KPI, or business intent, to increase the subscriber take rate, opened the door for AI to achieve heroic things! And it did. Netflix used AI to automate the personalization engine and show customers something they want to watch in under 90 seconds, which ultimately reduced churn. They did this by mapping a customer to a global community of subscribers who expressed interests similar to theirs. Their algorithms drove streaming hours up and lowered cancellation rates. Since the publication of their paper, the success numbers have probably improved even more: as their user base grows, they collect more data points on user behavior, which continuously strengthens the algorithm. 

Facebook revolutionized its business through AI. Starting in 2016, they released machine learning systems, which, today, are faster and more efficient due to the ever-increasing volumes of user data, and thus, opportunities for learning and the creation of intelligence. For example, they built an engine, DeepText, that understands text far beyond traditional natural language processing (NLP). This engine derives the meaning of text, and successfully generates leads by leading users to advertisers based on their conversations. Another example is the detection of bad content (e.g., violence, terrorism, SPAM). Facebook uses AI to detect fake or problematic accounts so that they can shut them down quickly and prevent them from damaging the credibility of the platform.

What if You're Not a FAANG Vendor?

Most of us aren't, but we can learn from the tremendous AI success stories they have provided the industry and start small. Whatever your industry, today's digital customer journey is hard to understand given its dependency on multiple systems, a mix of legacy applications and deep technical debt. Furthermore, operational silos within IT cause a lack of end-to-end transparency: obtaining the data necessary to understand the holistic customer journey becomes very challenging. In today's hyper-customer-focused era, this is a dangerous place to be. 

If you're typical of my enterprise customers, you're looking to shift your IT focus from "keeping the lights on" to providing IT services that drive business value....you'd love to come up with innovations like the ones at Facebook and Netflix! However, if you want to follow the AI successes of the FAANG vendor movement, you're probably faced with the epic hurdle of rewriting your key customer interfaces to collect the required telemetry.  A smart alternative solution is to instrument existing legacy applications to collect much of this data. I've done it with surprisingly low investment and risk (more information here). This provides your first stepping stone to greatness by enabling the businesses you support to learn from their existing customer usage patterns in real time. To get started you need to take the first step.  Identify a simple business intent, driven by an objective KPI (which is measured and predicted by modern AI), and just like that, you put AI to work to drive surprising business results.

 

Tell Me More!