Tom Laufer on AI-powered product analytics, contextual actionability, and lessons from his time at Google

Product State Q&A

Tom Laufer is Co-Founder and CEO at Loops. He was formerly the Head of Growth and Analytics for EMEA at Google
Website / LinkedIn

EC: What insights led to the development of Loops and what’s the story from idea to launch?

TL: Loops started from my own personal pain points. I was looking after growth and analytics for Google (driving Growth for products like Google Search, Youtube, Gmail, etc) and also worked as a Growth advisor for many companies. 

As we tried to identify the most impactful actionable insights to improve our KPIs, we had to look at thousands of dashboards, data sources, segments, cohorts, etc. Yes, we had great analysts and data scientists — we were flooded with data, great visualizations, but no insights.

We were struggling. 

It wasn’t just us.

From the beginning of the internet era, the general assumption in the industry was that we could empower data-driven decision-making by making data accessible. Let’s just put all the data in dashboards, and people would just understand where to focus on. 

Well, with the amount of data that is currently being collected, this assumption doesn’t work anymore. 

At Loops, we believe that out-of-context dashboards are not the solution, but the problem: You get to the office every morning and all these gazillion amount dashboards and data sources look the same. They are overwhelming, intimidating, inaccessible sometimes, and do not trigger you to take action.

In the long-lasting race to address the need to make data-driven decisions, we created another problem: The ability to translate all the data points into insight and then take an action.

As a result, product, growth and data managers are simply lost, failing to meet their OKRs.  

So the basic assumption behind Loops is that technology must empower professionals to focus on the most significant and actionable insights, not just landing them on dashboards. 

EC: What makes Loops more contextually actionable than other analytics tools in the market?

TL: Loops is focused on actionable insights — not just dashboard — and we do that by leveraging:

• Pushing the most significant insight — Each insight at Loops has a “potential impact” on your KPIs. Loops picks the biggest opportunities and enables you the perform sensitivity analyses to decide if you want to prioritize this insight or not. 

• Causation and not just correlation — Loops insights are based on causal-inference models. Loops basically simulates A/B tests, looking at users with similar intent and attributes to understand who causally drives your topline metrics. 

• Aligning with the company strategy and KPIs — The insights that are surfaced by Loops are 100% aligned with your top KPIs in a given quarter/year. 

• Semantic layer (Speaking your language) — Loops insights are in natural language, enabling the user to customize the definitions. It talks in features and not in raw events, in segments, and not in user properties. 

• Closing the loop and Improving over time — Loops measures the impact of every launch, be it with an A/B test or without. The causal models are improved based on the results of the actions taken and the learnings collected. 

Some examples of success stories from companies using Loops: 

• Optimizing your activation journey: Identify the features and journeys (“aha” and “habit” moments) that causally drive conversion or long-term retention. The team was able to change the first day experience and saw 20% lift in 2nd week retention. 

• Measuring the impact of a recent launch for a B2B PLG company: Measure the actual impact of a recent product launch for a company that doesn’t have enough traffic to run A/B test. The platform actually identified a negative impact of the launch, but thanks to Loops’ insights, the team was able to iterate on the learnings, launch another version, and see positive results on conversion. 

EC: What lessons did you learn during your time as Head of Growth and Analytics at Google?

TL: There are so many lessons. Here are three in the spirit of this conversations:

• It’s very easy to fall into this “data circus” phenomenon, a term by Elena Verna. We sometimes used to do analyses and dashboards that had zero impact on the business, they were ‘interesting to know.’ It took me some time to focus much more on actionability: Connecting every insight to the quarterly OKRs, accurately modeling the potential impact of the insight and doubling down on the story behind the insight, while adjusting to the different stakeholders. 

• It’s important to connect the marketing and product organizations. In B2C and PLG products the importance of creating one coherent user experience is crucial. The organizational structure sometimes limits our ability to align the user journey. Our biggest wins may happen when we align the marketing messaging with in-product experience (e.g onboarding), or when we leverage product data to improve the marketing strategy. 

• Don’t underestimate the learnings from failed experiments. We used to have a weekly email we sent out to everyone — where we celebrated wins (X% impact on retention for example). We didn’t share enough learnings from failed experiments, which, in my POV, really hurt our velocity and long-term success.

“Technology must empower professionals to focus on the most significant and actionable insights, not just landing them on dashboards.”
- Tom Laufer