Google’s Bihag Karnani On The 5 Best Ways to Drive Product Growth

As a part of this series, we had the pleasure to interview Bihag Karnani.

Bihag Karnani is a Senior Product Manager at Google where he builds AI personalization systems operating at Google’s scale — including features that determine what content surfaces for whom inside Google’s AI Overviews and AI Mode.

Thank you so much for joining us in this interview series! Before diving in, our readers would love to learn more about you. Can you tell us a little about yourself?

I’m Bihag Karnani, I am currently working as a Senior Product Manager at Google on some of the most critical issues in Artificial Intelligence today. Specifically, I focus on how AI systems make decisions about what content to display to who and how personalization affects what individuals see when asking AI a question.

My role is central to both of these efforts through AI Overviews and AI Mode; and I have been responsible for developing many of the key features that control how subscribed source preferences (such as preferred publishers) and personalizing elements (and other signals) impact what an individual user sees when searching at scale. The decision making process of the product decisions my team is making around identity, preference, and trust will be the quiet determinant of whose content is seen, by whom, and when due to the generative search that will alter the experience of over hundreds of millions of people discovering new information.

Prior to this position, I developed much of the infrastructure that supports this effort including developing and leading the Subscription Linking API which establishes links between publisher/subscriber identities and Google account identities and developing Reader Revenue Manager, a platform that has enabled 6,000+ content publishers worldwide to generate tens of millions of dollars in revenue via AI and Machine Learning that converts casual readers to paid subscribers. I was also the leader of Google’s Partner Hub Platform providing AI powered personalization solutions to 108,000+ global partners at Google Cloud Next ‘24.

Prior to Google, I led Smart Home IoT product portfolios for Canadian Tire Corporation. I have also served in Growth Strategy positions for Open Text and Enquero where I worked with large technology companies such as Cisco Systems, VMware and Intel. My first job out of school was in Quality Engineering at Samsung which gave me an appreciation for building products that really do work at scale.

I have an MBA from Purdue University, but my most valuable education came from seeing how real product decisions are made — those where the numbers say one thing, your gut says something else, and then you pick.

What led you to this specific career path?

My path in my career was more about making several smaller, yet important decisions rather than taking a direct route.

Firstly, I remember the moment at Samsung (where I worked as an Engineer in R&D), I did well with the technical side of things; however, the thing that gave me energy the most was not the technical aspect, but the point where technology solved a business problem. The question I asked myself continually while at Samsung was “why are we doing this?” Rather than asking “how,” which ultimately led me to pursue a Business degree. After completing my MBA at Purdue University, I went to Genpact as a Management Consultant and was immediately immersed in the B2B space (i.e., solving operational and/or revenue challenges) for clients such as Cisco and VMware. Although I enjoyed the process of solving problems, there were some challenges in every case. Every time I would be able to identify the issue(s) and recommend ways to solve these issues, I could never create solutions. In other words, I was always removed by one level from creating solutions. This experience provided clarity for me. While I was happy to provide advice on products, I truly desired to develop products.

This realization steered me into product development. I began my product development experience at OpenText where I developed expertise in developing products related to B2B sales focused products and learning how enterprise software actually gets used (or does not get used). Following that experience, I moved to Canadian Tire Corporation (in Toronto) where I was involved in creating a new product from scratch (assembling a team and launching the NOMA iQ line of smart home products for the Canadian market). One of the key aspects of this experience was gaining insight into what it means to be responsible for a product throughout its entire life cycle (from concept to delivery and into customers’ hands).

Finally, I arrived at Google. At Google, all of the knowledge/experience I gained during my tenure at B2B platforms, consumer products and growing strategies finally aligned. As part of this journey, I have created various products including Reader Revenue Manager for 6,000 news publishers, developed the Partner Hub platform for 108,000 global partners and currently working on AI personalization within Google’s AI Overview/Mode — each of the roles has enhanced the capabilities developed in the prior role.

In summary, I believe I have been drawn to being a product manager based upon continuing to follow what gives me energy until I reached a juncture in my career where all those elements combined.

Can you share the most exciting story that has happened to you since you began at your company?

One of the most exciting things I’ve done is to recently launch customization and personalization of user content preferences on AI Mode, AI Overviews and Gemini. This is a feature where when a user indicates to us their content source preferences or links a content subscription account to their Google account, AI Mode, AI Overviews and Gemini pay special attention to these sources and highlight to the user when their responses are curated through the user’s source preferences. This has been shown to improve user trust in the AI responses and also lead to more traffic to the content creators. Why this was so exciting for me was that I had always hoped but didn’t really know if I’ll ever be able to work on Google Search — the pinnacle of technology that has influenced humanity at a scale no other technology has. I still can’t believe this is what I work on now.

What are some of the most interesting or exciting projects you are working on now? How do you think that might help people?

I’m working to expand the kind of customization of user experiences on AI Mode and AI Overviews I explained above. So there’s more options that we’ll provide users to get more value out of AI forward surfaces, not just Search but also across Google’s portfolio of AI forward products. I think this will help people in two ways. For readers, it means AI surfaces that actually know what you care about and who you trust. For content creators, it means their most loyal readers find their content even as the discovery surface shifts from keyword search to AI-generated answers. That’s a problem worth solving.

You’re a successful business leader. What are three traits about yourself that you feel helped fuel your success? Can you share a story or example for each?

Grit — This is the one characteristic I think beats almost everything else. I believe that life will throw its challenges at you, and only those who continue to persevere regardless of how tough it gets will succeed and move forward. Within Google, one of the toughest problems we have faced is to launch one of my products, Reader Revenue Manager, to India given India’s unique landscape of legal, regulatory and market nuances. 3 of my predecessors including my manager tried and failed. When I joined the team, I found collaborators who were passionate about bringing the product to India, and created a very unique multi-phase strategy where we only launched 80% of the product features to the country first, and continued to get the required permits to launch the rest of the 20% as we continued onboarding content creators to the platform. The product was an instant hit with over a 1,000 content publications onboarding in less than 3 months with nearly no marketing efforts.

Growth Mindset — I’ve always spent a lot of time, energy and resources on continuously improving myself and my craft.I had to learn a lot of product approaches to be able to reach the level that is acceptable to Google, but once I joined Google, I realized that there’s so much more that needs to be learned. I had to learn to give very difficult feedback very effectively so it only builds the relationships further, to negotiate with ease and walk away from the table with all parties feeling good about it. Once I realized that I needed to learn these skills, I pursued courses from Crucial Conversations, read the book Never Split the Difference and then pursued Chris Voss’ Masterclass to further cement my understanding. I recently completed courses on Product Taste, Product Sense, Executive Presence, Influence and Machine Learning, all in this calendar year. I believe that in a knowledge economy where learning is significantly accelerated with AI, one has to invest in themselves to stay relevant and to stay ahead.

Do you have any mentors or experiences that have particularly influenced you?

Two experiences have impacted me a lot -

The first experience was when I started my own business (Magic Technologies) in 2016. I created a Software-as-a-Service (SaaS) platform for restaurants that helped them establish an online presence using a variety of tools such as websites, apps, point-of-sale systems, and kitchen management software. In less than six months, our customer base increased by 800%. This experience provided me with the opportunity to be completely responsible for decisions and consequences of those decisions. With this type of responsibility you understand why things worked or did not work. There is no other party to blame. I learned more about growth, priority setting, and the costs associated with switching between contexts during these eighteen months than I could have ever learned while working in a corporate environment creating products. Since then, I strive to allow each member of the product teams I lead to feel ownership over their desired outcome rather than only their assigned task; this is based on my knowledge of how owning an outcome will change how they approach their responsibilities.

The second experience was realizing that the greatest product managers are not those who bring the best ideas into meetings. The greatest product managers are the ones who set up environments that support bringing the best ideas into meetings. At the beginning of my career, I believed that being a great product manager meant having a stronger instinct and better answers than anyone else in the meeting. What I have discovered is that it really means developing the skills needed to ask better questions, remove barriers that prevent teams from moving forward, and continue to develop curiosity in order to identify a good idea regardless of where it originates. The transition — from being the smartest person in the room to making the entire room smarter — has been the largest transformation I have experienced as a leader, and it continues to be something I work on daily.

What have been the most effective tactics your organization has used to accelerate product growth?

The two tactics with the largest impact have been what I refer to as “Identity-First Growth” (building an infrastructure that connects a user’s identity to their experiences) followed by “Partner-Led Growth”.

Google’s Subscription Linking API established an identity layer connecting Publisher Subscriber Identities to Google Accounts. This Identity Layer was the basis upon which all subsequent Personalization & Growth Features were developed. Once we knew who a person was (what they subscribed to; what they liked; level of engagement), we could develop and implement growth strategies that would truly affect behavior — instead of simply tracking it.

The second most impactful strategy was developing “partner-led growth.” With over 108,000 global partners on the Partner Hub platform, we quickly discovered that if we could simplify success for our partners, then those partners would be our greatest growth driver. Therefore, each dollar invested in making it easier for our partners to reduce time to value — and ultimately generate revenue — led directly to partner advocacy and the overall growth of the ecosystem through which no amount of advertising or promotion could drive.

What do you see as the biggest challenge with respect to scaling a product-led business?

Lack of experimentation. Every product has a unique user set and audience that has it’s own unique requirements. It’s crucial for product led businesses to keep engaging in small experiments to understand their audience’s preferences and needs better to continuously improve the product. A good practitioner will not just look at experimenting in the product feature or experience that isn’t delivering results, but also experiment in areas that are already doing well to see if they can do better. Duolingo’s product testing first growth approach is a case study I highly encourage people look into to learn from.

What, in your view, is a good litmus test to screen for a skilled and effective growth manager?

The only question I am asking is: ‘Tell me about a time when your company ran an experiment that failed and how you reacted to the failure.’

All successful growth professionals have had some experiments fail. However, skilled growth professionals know precisely why they were optimistic the experiment would be successful; they can describe exactly what the results indicated, and they will describe changes that were made due to the failure. To them, the failure of the experiment was nothing but data — nothing more than data; they never saw it as a problem.

When I hear an answer with specific details (this was our hypothesis, this was our methodology, these were the results of the data, this is what we changed) I see evidence of someone who has worked through the process; whereas if I get a vague response such as “we learned something from failure” I believe that person has merely read the correct books on growth. Growth is an empirical science. Therefore, those growth managers that succeed are people who are naturally curious about why certain things function, rather than simply content when things function.”

Can you describe a product growth tactic you or your team has used that was more effective than you anticipated? What was the goal, how did you execute, and what was the outcome?

The tactic that surprised me most was opening up Reader Revenue Manager to small independent creators and publishers on WordPress through our Sitekit integration.

The original goal was straightforward and modest — we wanted to extend RRM’s reach beyond large enterprise news publishers, who had been our primary focus, to smaller independent creators who were building subscription businesses on WordPress. Sitekit is Google’s official WordPress plugin, already installed by millions of site owners globally, so the integration made technical sense as a distribution channel. We treated it initially as a long-tail experiment, not a core growth driver.

The execution involved building a simplified onboarding path through Sitekit that allowed WordPress publishers to activate Reader Revenue Manager features without the technical lift that had previously been a barrier for smaller operators. Where enterprise publishers had dedicated engineering teams to handle integration, independent creators needed a near-zero-friction setup experience. We rebuilt the activation flow with that constraint at the center.

What we didn’t anticipate was the scale of latent demand sitting in the WordPress ecosystem. Independent creators who had been building audiences for years but had no viable monetization infrastructure adopted the product rapidly. The outcome exceeded every internal projection — the Sitekit channel ended up accounting for 70% of our total publisher volume for 2025.

The lesson I took from it applies broadly to product-led growth: the most significant growth opportunities are often not in moving upmarket toward larger and more complex customers, but in dramatically lowering the floor so that an entirely underserved segment can access value they couldn’t reach before. We weren’t building a new product. We were removing the barrier that had been invisibly excluding most of the market all along.

Thank you for all of that. Here is the main question of our interview. Based on your experience, what are your “5 Best Ways to Drive Product Growth”? If you can, please share a story or an example for each.

1. Build within existing ecosystems — leverage native habits as your distribution engine.

During my time leading the NOMA iQ smart home IoT portfolio at Canadian Tire, we recognized that our users weren’t starting from scratch; they were already deeply embedded in established ecosystems like Google Home, Alexa, and Apple HomeKit. These platforms served as the central nervous system of their domestic lives. We realized that for any new hardware to succeed, it couldn’t just exist in isolation; it had to thrive within the infrastructure users already trusted.

Instead of forcing users into a proprietary app — adding friction to an already complex setup — we prioritized native integration with the major home ecosystems. This shifted our strategy from competing against these platforms to being distributed through them. By meeting users where they already were, we gained immediate entry into their daily routines without requiring them to learn new behaviors.

This approach dramatically shortened the distance from the point of purchase to the “aha” moment of first successful control. The results were measurable: we saw a significant surge in product adoption and activation, coupled with a substantial decrease in customer support volume and product returns. When you reduce the cognitive load of adoption, you increase the likelihood of long-term retention.

The broader growth lesson is clear: identify where your customers already reside and build the bridges that allow your product to flow naturally into their existing workflows on their own terms.

2. Organize your growth strategy around sustainable indicators versus vanity metrics.

Growth teams typically fall victim to metrics that can be quickly changed — installs, sign ups, etc., vs. metrics that indicate long term retention and revenue. Although metrics that indicate short term success may look like advancement when the actual product is losing customers acquired via the initial metric.

I developed frameworks for measuring real-time propensity and engagement signals across all of our publisher bases while working at Google. For example, a publisher signed-up for Reader Revenue Manager, completed on boarding and went silent for thirty days; there was no first meaningful value moment achieved until it did. The acquisition was successful; however the activation failed. Furthermore, since the dashboard reported the publisher as having completed the sign-up process (and therefore a successful acquisition) and not a ghost, we did not know about the issue.

The answer lay in defining the specific behaviors that identified engaged users from dormant ones — whether publishers had linked their payment processors, created their first subscription offering and experienced at least one conversion event in less than two weeks. Those three behaviors were much better predictors of 90 day retention than any of the top funnel metrics. Once we understood that, we were able to develop targeted solutions that assisted publishers in moving past those specific gates. Long term growth will come from not merely acquiring customers but from the disciplined effort of converting those acquisitions into true engaged users.

3. Create growth a platform function not a feature

The most scalable approach I’ve ever seen to growth is creating growth not as functions you build once but as continuous improvements in platform functionality that you provide every product with. If you treat growth as a function then each new group builds the exact same mechanics again. However if you create growth as a platform, each new product has access to a growing body of growth functionalities it wouldn’t have otherwise had.

At Google, when we introduced Subscription Linking API — linking publisher subscriber identities to Google Accounts — we weren’t developing it for one product. Rather we made it a foundational mechanism enabling subscription signaling across multiple surfaces including Search, Discover, AI Overviews and AI Mode. When we later launched Preferred Source Signals in AI Overviews — we didn’t establish another identity layer; we leveraged one that existed.

Prior to designing your next growth function ask yourself if you’re developing something that solves once or something that creates opportunities for all subsequent growth initiatives to be more effective. Teams that win in the next decade won’t be building the most unique individual features; they’ll be building the platforms that enable features to be unique.

4. Obsess over “aha” moments — engineer the shortest path to them

Each product has a point at which a user finally understands why they bought the product — where the promise of value converts to reality for the first time. While many product organizations understand this point exists in some form, few organizations define it precisely; measure how long it takes for new users to arrive at this point; or deliberately design the path from here-to-there.

While working on Canadian Tire’s smart home IoT portfolio — the ‘aha’ moment was unambiguous: the first time a user successfully controlled a device from their mobile phone. All experiences prior to this moment represent friction. All experiences after this moment represent momentum. As we analyzed our data — a large portion of our users who bought devices never arrived at that first control event — not because the product itself failed; but because the configuration experience lasted longer than expected.

We re-architected our entire onboarding experience with a singular focus on this moment as our North Star — evaluating each screen based solely on its ability to either shorten or lengthen the distance from completing this moment. Through this approach we saw an increase in activation and a 35% decrease in customer service calls. My recommendation: write your ‘aha’ moment as a single sentence, determine how long the average user takes to achieve this moment — then make it your mission to reduce that time

5. Acquisition alone doesn’t grow your company — it grows the treadmill you’re running on.

Acquiring users while simultaneously losing as many is not “growth”, it’s a waste of time. Churned customers are both lost revenue and the cost of acquiring (and therefore losing) another one.

I’ve found that no other investment in growth has paid such consistent dividends than to stop the leak (of churned customers) before increasing the flow (of acquired customers).

When I worked for Genpact (inside Cisco’s subscription and licensing division), our team identified that the reason for loss of automatic renewals was due to the complexity of the renewal process. Channel partners would lose interest in completing the renewal workflow prior to the final step, therefore we did not implement an additional marketing campaign to promote the renewal. Instead, we simplified the renewal workflow to remove the steps where channel partners typically lost interest. The results were a $6 million dollar annual increase in automatic renewals — solely based on preventing users from churning. Not by adding additional users.

At least half of all churn is preventable — and preventable at a small fraction of the cost of replacing those customers through acquisition.

What is the number one mistake you see product marketers make that may actually be hurting their growth outcomes?

Product marketers are generally optimized for acquiring users (i.e., awareness, clicks, sign-ups), which are typically easier to tie back to a specific action taken by a marketer. However, the greatest potential point of leverage for marketers in terms of how users interact with a product is at the very beginning of that interaction — i.e., the moment in time where the user experiences the primary value proposition offered by the product. If that value proposition cannot be delivered early enough, then all money spent on acquiring users will have been wasted. I can give you numerous examples of this happening. A publisher signs up for RRManager.com, completes on-boarding, and does nothing for thirty days after signing up for RRManager.com because the first substantive value experience — e.g., the first subscriber conversion generated by RRManager.com — took too long to occur, or was too difficult to accomplish. Acquisition occurred; Activation failed. And since the dashboard for the marketing team simply shows “sign-up” and “activated,” the marketing team has no idea that there is an issue.

To solve this issue requires a surprisingly straightforward solution: First, define exactly what constitutes an activation event. Second, measure activation events obsessively. Third, view the time path from sign-up through activation as a product issue, not a marketing issue.

It has been said that our mistakes can sometimes be our greatest teachers. Can you share a story about the funniest mistake you made when you were first starting? Can you tell us what lesson you learned from that?

Early in my career at Samsung, i was responsible for building daily status reports that went to senior leadership across multiple geographies. I had created what i thought was a great daily status report; it was beautifully formatted with color coding and conditional formatting and an elegant summary dashboard.

What I did not do before I sent the report out to the various regional locations is check their date format setting by time zone. This caused the dates on my report to display differently based on whether or not you were viewing it in North America vs. Asia — so what I intended as January 3rd showed up as March 1st to half of the leadership team.

A feature launch I reported as on schedule for January looked like it was already three months overdue to half the leadership team.

I spent two hours on an emergency call explaining to very senior people why the feature was not three months late, and yes, I understood why they thought the report indicated that the shipment was three months late.

The lesson I took from that experience, and have applied ever since — is that the most dangerous assumption in product work is that your user will see/experience what you intend. I built the report for how I read it. I never even considered how it would render for someone else.

That instinct to test for the unexpected experience has been fundamental to how I approach product design. Every feature I’ve shipped since has a version of the question “what does this look like to someone who isn’t me?

We are very blessed that very prominent leaders read this column. Is there a person in the world or in the US with whom you would love to have a private breakfast or lunch, and why? He or she might just see this if we tag them :-)

I would love to take Ben Thompson of the Stratechery blog out to dinner. I’ve learned more from reading his publication that I have in my MBA and engineering degrees combined. His perspectives on what is going on in the tech industry, what are things that truly matter and explaining these perspectives simply are awe inspiring characteristics I aspire to grasp.

Thank you so much for this. This was very inspirational, and we wish you only continued success!