Playbook
GTM Playbooks
How GitGuardian Leverages Data and GitHub to Fuel Its Growth
Read more
This article is part of François Dufour’s “New Category or not?” series on positioning and category definitions in B2B software. There, founders and CMOs reflect on their positioning and category-naming journeys and decisions. They share advice for those navigating such projects.
Jen Grant took over as Looker’s CMO in 2015 when Looker had 130 employees and right after their series B. Looker already had a lot of happy customers. Yet, there was no clear description of what Looker was on its website. No clear positioning even internally.
Jen had to change that. This led to having to answer one of the first key questions in positioning and messaging: was Looker in a new category, a player in an existing category, or something in between? Her experience in this market brought about many learnings which are shaping her thinking today as COO of Cube, the open source and unified semantic layer company. Jen walked me through her process, decisions and learnings of that time and how it impacts her strategy today, as the data analytics market is once again at an inflection point.
First, Jen had to understand how Looker really worked, its benefits and how it was really different - or similar - to other solutions for what really mattered to customers.
For any incoming head of Marketing or Sales, truly grasping that is not easy. Especially when you are not in the target audience and not using the product for your own sake.
Jen first went to engineers to figure that out.
“They would talk about the semantic layer, get into the details of LookML - SQL except better - that we had this amazing trademark on aggregations. Etc. Okay, I said, so how does that help a human?”.
Sales Engineers helped her understand the problem. They could really translate for her from features to benefits and define the issues that the customers were having. It’s not surprising as they talk to customers daily AND get the technical aspects of the product.
So that was a critical place to start. Jen had done that as well at Elastic. It’s common for Sales Engineering to become a Marketer’s - especially a Product Marketer’s - best friend in the early days.
Thanks to that, Jen then figured out an essential challenge created by Tableau: every Tableau user was defining their own data model.
“I realized that Tableau was creating a very real issue where everybody fights because their data is different. We decided to call that “data chaos” which happens because of the data silos that Tableau had created. That really helped us describe the value to the business of Looker’s data model layer and LookML that can unify metrics and their definitions.” Interestingly enough, she states, history repeats itself: this "data chaos" problem is being exasperated by the additional new data sources and new data consumers in the modern infra world and their increasing need to build or use great data apps today!
When Jen started to work on positioning and describing what Looker was, that question arose very quickly: was this a new category? Looker considered the following criteria to inform that decision:
Jen and the team realized that the category was very strong. All of the leads - organic and paid - coming to the sales team were people looking for a BI solution. And the difference Looker made was important but not fundamentally changing it.
They concluded that it would be very likely. The sales team even tried to share a different category with prospects, who ended up saying, “Oh, like BI right?” Defining their own category would have confused people at the time.
Looker differed from the main category players at the time because they did not have a database (arguably they do now that they got acquired by Google and therefore are tight with BigQuery). They sat on top of whatever DB one had and that was new. Jen felt that avoiding the lift and shift requirement that all BI players had was a benefit they could explain and highlight even if they were a player in the BI category. As well as their benefit of being a better BI that didn’t create data chaos.
This was important for Looker but not critical since they were not targeting the Enterprise much initially. They also thought that the role of analysts was not as important as eight years earlier when Jen was at Box.
Note that companies often confuse category creation with Gartner creating a magic quadrant for the category the company is championing. That is a mistake, even if you solely target the enterprise. Gartner will create an MQ once and only once they receive enough inquiries about your space, you, and your competitors to justify creating and maintaining an MQ. So play your game first and create your category - or niche - in the eyes of enough customers first and make sure analysts know you, your approach and momentum.
They asked themselves:
“Do we have a bullhorn? How much are we ready to spend on category creation and what is our patience for that?”
Unlike what Jen had known at Box with Aaron Levie, Looker did not have a CEO who loved being “out there’ talking to the press, keynoting at conferences, and being super active on social media.
Nor did they have massive Demand Gen or Brand budgets initially that would power their Demand Gen efforts. It became obvious that their Demand Gen engine had to latch on to what people knew and that they should latch on to BI and be in that consideration set.
“We wanted to show up in search results when people were looking for BI tools”
As a result, you guessed it by now, Looker stayed away from defining a new category and instead expressed how they were better than existing BI tools. That’s a decision that Jen and the Looker team did not regret.
What was the alternative? Looker could have defined their category as something along the lines of the Unified Semantic Layer with visualization on top. At the time they called that the Data Model Layer, which is basically LookML. At that point, they stepped back and wondered: “Okay, are we going to create a new category above and beyond BI?”
They saw that Looker was solving two big issues: the data chaos (different teams using different definitions of the data) and the data queue (having to wait for the data gatekeeper to get to your request for new reports). By creating this unified semantic layer and offering great visualization options, they were solving both. But they realized that the pull of the BI category was too strong and Jen assessed that the market was not ready for that conversation. She believes it is today (and so do we at Decibel! That’s why we invested in Cube), but also believes the Looker team made the right decision back then.
Tableau was the BI category leader of the new wave of BI tools and in analysts’ good favors. It had a big headstart with them and that proved frustrating for Looker.
Jen recalls how the analysts did not really get Looker’s differentiators initially, convinced Tableau truly did everything it claimed.
Looker’s big value proposition was sitting on top of any database and creating a unified data model so that everybody in an organization would look at the same metrics and the same definitions. But analysts would retort “well, that's no different from Tableau”.
“I would bring in the Head of Product, the founder, and other technical people to say “ No, Tableau does not sit on top of a database, it moves data from wherever it is into Tableau to do the analysis. That's what the workbook is about. That's why the data model is created in these different workbooks”. And the analysts always said, “Oh, but Tableau says they can do SQL queries on top of any database.” But just because they say it doesn't mean it's happening. Depressing”.
Jen recalls a Board Meeting when Rob Ward from Meritech Capital asked: “How are we doing with Analysts?” Jen candidly answered “We are doing terribly! Tableau did a great job and this is - and will continue to be - an uphill battle.” He laughed because I was honest and he agreed.
Was that really a hindrance for Looker? Jen admits that it was but only for big enterprise deals.
Her efforts paid off and Looker eventually got into an MQ but was labeled as niche. They finally changed their perception by getting enough customers in front of them, who were big enterprises, who explained the differentiators.
At some point - admittedly for a short period of a few months, Looker wanted to become more of a developer platform. So they started talking about themselves as a platform that developers would build on top of.
They updated their homepage, which looked a little more cartooney and Docker-like. Most critically, they changed their site navigation and, at the top, removed the Product tab and kept only a Platform tab.
“That TANKED our website leads. We realized that everybody was looking for BI! Platform makes it seem too advanced - not a product to help business users. If it doesn’t say Product, business users will turn away”.
So they brought Product back into the nav, which lived alongside Platform.
They learned: “Let’s not get fancy. We need to be accessible”.
Eventually, toward the end of Jen’s tenure, Looker tried to go past BI to say “this is Data Analytics Platform that doesn’t just replace BI, but also your Mkg Analytics, Sales, etc.
It could have been their category at that point but it was too early back then. Now Looker uses “Big Data Analytics Platform”. Jen supposes that, with Google buying Looker, there really is a DB behind it - BigQuery - which allows them to claim that. Even though they still say they are multicloud…
There are many reasons why Jen just decided to join Cube as their COO. Cube is creating the new category of Universal Semantic Layer, which she and the Looker team had considered creating a few years ago.
First, there is even more data created, with more data sources, CDWs, data lakes, lakehouses, and more tools to create, extract, load, transform, analyze, and consume all that data. Therefore there is more data chaos, more data bottlenecks, and more demand to be able to exploit that data securely and consistently, in great data experiences and apps.
Second, Cube does have a bullhorn: its enthusiastic and growing open source community (7.5k members in its Slack community in Feb 2023). Community-led and PLG are strong forces that shape, drive adoption, and evangelize new categories without requiring big brand and demand gen budgets.
Third, the semantic layer as a concept is much better understood by many data leaders these days, as highlighted in a recent discussion at a Decibel Data Council that brought together Chief Data Officers and other data leaders. That Semantic Layer, which seemed to be a pipe dream a few years ago to them, is now within reach and, in some organizations, already firm in the grasp of their hands.
Fourth, as Jen shared in another discussion about category creation:
“You do not want to be the only one creating a new category. If it’s just your company, it’s not really a category. ”
However, the Universal Semantic Layer category is clearly tempting other companies from different parts of the data stack and they are building products to address it. Recently, dbt Labs put aside its nascent semantic layer product to buy Transform with its early-stage metrics layer tech and are hoping to shore up its ability to sell to the semantic layer market - a clear validation of the category. Cube, of course, has long been in this market and has established a significant lead with security, authentication and caching layers, as well as their ability to work with multiple data sources.
Thanks a lot, Jen for sharing this journey with category definition and positioning Looker. I really appreciate your candor and, as always, great sense of humor and directness.
.png)
