Rao Talasila most recently worked at Xnor.ai, which was acquired earlier this year by Apple for about $200M.  Rao has expertise in a variety of areas, but his main focus at Xnor was helping commercialize their technology.  Before Xnor, Rao worked at a data science company (Mu Sigma) where he helped build their Asian business from the ground up.  Before that he was a Director of Engineering at Cisco Systems.  So, Rao has expertise on both the technical and business side, but the key values he brings to startups are:

• Technology commercialization – how to take technology that’s been developed and move it to commercialization most efficiently
• Achieving product / market fit
• Client development

What are strategic framework/ critical considerations for setting up a data science capability as a startup?

Data driven decision making is becoming a must-have capability for startups and enterprises alike in this decade and beyond. Answering the below three key questions can help a startup in setting up the right Data Science capability:

What are the key business insights the startup wants to derive every day/week/month? In other words, as a CEO of a startup what are the key metrics I want to track in my business periodically?

What kind of data is needed to support the derivation of these business insights? How much data does the startup already have, and what are the additional data that needs to be added from third parties?

What kind of Data Science team that needs to be built to realize #1 & #2 above? How much can this be done in-house vs outsource?

Are there some unique things people don’t talk about when approaching product/market fit?

The product/market fit is usually a discovery process, and not a one-time event. Talking to potential customers, willingness to quickly test out hypotheses, and fast-fail approaches is how one can find optimum ways to find the product/market fit. 

One of the things that people often ignore is how luck & timing plays an important role, namely, being at the right time and right place. I have personal experience in a couple of startups where our product/market fit worked very well for a bunch of customers, but we were too early for the market takeoff.

What are the key similarities/differences of focusing on US vs Asian business we should take into consideration as we plan to globalize our companies?

The U.S. has a high-trust society with a well-developed legal system, while Asia in general has a low-trust system with evolving legal infrastructure. This manifests into the following business environments:

• In the U.S. customers would be willing to try out a startup’s products just based on the strength of the products benefits

• In Asia, customers want to make sure that the startup is “trustable”. A well-established brand name from the U.S. coming to Asia would gain some automatic trust. For the rest, partnering with a local company in Asia is one way of building the trust. In countries such as China, South East Asia & India there is a large mobile-savvy customer base that can propel a startup’s business to exponential growth with the right product/market fit. This means that a startup needs to know how to handle the deluge of demand, and keeping ahead of the copycats.

Could you share few stories of how Xnor.ai went about achieving their product market fit? How did Xnor.ai go about narrowing their target audience and go about client development?

Like any other startup, Xnor.ai worked with many customers in various verticals to find the product/market fit. The verticals were really diverse and it included Automobile, Aviation, Consumer Electronics, Entertainment, Government, Oil & Gas, Agritech and many more. 

Edge AI (AI on the edge devices such as IoT devices, Drones, Smart Home Cameras), the domain of Xnor.ai, is in itself an emerging area and many customers are themselves trying to figure out how they can derive business value out of Edge AI. Below are the three stages that customers cycled through during their journey with Xnor:

Stage 1: Customers have some pain points to solve and are interested to explore incorporating Xnor’s AI models . There are also some customers who are curious about Xnor’s technology and want to kick the tires so that they don’t miss out the boat on Edge AI

Stage 2: Customers sign up for a Proof of Concept (PoC) with Xnor

Stage 3: Customers develop a strong conviction after the PoC that Xnor’s technology can help solve their business problem. Xnor builds models that would solve their specific business problem, and also demonstrates the commercial viability at scale

In the initial days of Xnor, much of the activity was spent on Stage 1 & 2. Our goal at Xnor was to move customers quickly to Stage 3. During the process of cycling customers from Stage 1 to 3, our business model and pricing went through more than a dozen iterations. Basically, the more our customers are in Stage 3, it is more of a reflection on how good we are able to hone in our ability to find the product/market fit.

For example, we worked with a couple of well-known global automobile brands. They absolutely loved the PoC that we built for them, but we had a long waiting period to get into Stage 3 with them because the automobile industry has a 2-3 product development lifecycle.

In another example, we worked with a Smart Home Automation company where the customer’s evolution from Stage 2 to Stage 3 was the fastest as the customer was keen in working with Xnor to help them dramatically reduce their operational costs.

What are some mistakes that you saw early in your career that you have avoided? On the flip-side of that question, how did you guide your actions to ultimately be so successful? What was the guiding set of principles or beliefs that kept you focused or in the good graces of success? How did you do it?

In my early days of career, I was enamored with my own visions of technology and why a customer should use the product that our company built. When I became a key decision maker in a company, I carried this philosophy and ended up with sub-optimal results. I learnt a lot from such battle scars and became a firm believer in the adage “you learn much more from your battle scars and failures”.

The lesson that I learnt from my previous mistakes is to focus on the customer’s context and why the customer should buy our product/service. In other words, think from the customer’s perspective. This sounds like a simple prescription but needs a lot of discipline and humility to practice.

Is there a methodology you use to figure out how to commercialize a product?

I try to answer the following four questions during the product commercialization process. Every interaction with customers needs a crystal clear approach to elicit the responses for these questions. 

What is the product the startup is providing?

1. Who is the customer for this product? 
2. Why should the customer buy this product?
3. What is the price the customer is willing to pay for this product?

The answers to these questions would change as you get closer to a practical product/market fit, and from the feedback gained talking and working with potential customers.

What are the the top two tips you can give for client development?

There are no silver bullets that would work all the time, we have to adapt to the constantly changing market conditions. From my personal experience, I know how challenging it is to do a startup. One day you feel you are at the top of the mountain, and the very next day you are wondering how you got into this mess. But don’t worry. This is normal! 

If I have to give just two tips for client development, it would be:

1. Go and talk to customers, meet them on their turf (post-Covid lockdowns), learn from them, think from their perspective
2. Don’t be afraid to put milestones with customers. For example, if we deliver this product to you (Customer) by 60-days, will you buy our product? Can you send us an email or Letter of Intent for buying our product?