“Whether you’re talking to other founders or looking up books and blogs, they teach you the wrong thing when it comes to vertical SaaS. They made me commit a lot of mistakes. Even investors are not able to give you the proper advice because they’re usually working on horizontal SaaS… We are now a billion-dollar-plus-valued company, but it came with a lot of battle wounds,” says Guru Hariharan, founder and CEO of CommerceIQ.
The company, based in Palo Alto, Seattle, and Bangalore, helps retail brands with algorithmic ecommerce on channels like Amazon. Hariharan calls it the Shopify of ecommerce. While Shopify helps brands build an online presence, CommerceIQ helps them interface algorithmically with ecommerce marketplaces, where 90% of their sales are happening.
Algorithms determine everything on a site like Amazon. So brands also need automation with SaaS products to make ecommerce algorithms work to their advantage. In the old days, they would deploy human managers to build relationships with physical retailers, jostling for shelf space or figuring out pricing. Now algorithms decide shelf positioning, as in showing up in search results, and competitive pricing, which can keep shifting. “And you can’t take an algorithm out to dinner,” quips Hariharan. You let a SaaS product with an AI brain talk to the algorithm.
Hariharan gravitated towards machine learning during his MS at the University of Texas, Austin, and a stint at IBM Research. “It attracted me because of the foundation of math and the amount of stuff you can do with data that can be very impactful.”
He earned his spurs on the practical applications of ML at Amazon, where his team automated vendor management. “For the first time in the history of retailing, we replaced people with machines on the retailer side to work with the millions of vendors you need to build the Earth’s largest selection of products. It allowed Amazon to remove the human capital bottleneck in its expansion to tens of millions or even 100 million plus vendor and seller entities.”
It was during that period in the late noughties, as the head of the Amazon Dev Center at Phoenix and manager of Amazon Seller Services, that he had an epiphany. “Amazon was becoming an algorithm that other retailers would have to compete with. And in any human vs. algorithm contest, it’s hard for the human to win.”
That led to his first startup, Boomerang Commerce, which built a dynamic pricing engine to help other retailers compete with Amazon. Seven years later in 2019, he sold Boomerang’s retail analytics platform to Lowe’s. He then focused on Boomerang’s offshoot CommerceIQ which became a separate business to help brands deal with the algorithms on the retailer side. Three years later it became a unicorn with a $115 million fund raise.
Hariharan learned his lessons the hard way in Boomerang Commerce and tried not to repeat those mistakes in CommerceIQ. Now, for the first time, he shares his hard-earned playbook for vertical SaaS, which he says has to be fundamentally different from what you use in horizontal SaaS. Let’s dive into his three most important lessons.
Lesson 1/ Don’t believe the usual PMF definition
“In a vertical market, you tend to be a domain expert and you can be fooled by the revenue you get because there can be a lot of services underneath which are not repeatable. It’s actually not very difficult to get a million dollars in vertical SaaS,” says Hariharan. “We got fooled by that. And just measuring traction by revenue and throwing fuel to the fire almost bankrupted me in my previous company.”
Not understanding the true PMF (product-market fit) can lead a vertical SaaS founder down the garden path after getting to $1 million in revenue, often spurred on by investors to ramp up marketing, hire more sales folk, and so on. Instead, you have to stay focused on finding the real PMF. That was the big learning. It’s different for a horizontal SaaS product applicable in multiple domains, where $1 million ARR can be a signal for blitzscaling.
“A term you hear a lot in CommerceIQ is C3. Whenever we bring a new product in, we just think about three successful customers, not revenue,’’ says Hariharan.
The milestone he set for CommerceIQ in the first leg of its journey was C20 – that is, 20 successful customers. “Please don’t hold me accountable for revenue in the first year,” he told the board.
This can be jarring in a board meeting where a revenue milestone is usually the first point on the agenda, but Hariharan stuck to his guns, mindful of the scars from Boomerang. “We actually trained ourselves internally in the company, and the board, that until we hit C20, nobody’s gonna talk about revenue, and that was very hard,” says Hariharan. “We maniacally focused on C20 and had a goal of getting to it in 18 months. We got there in 15 months.”
Lesson 2/ Don’t miss the trees for the forest
“Missing the forest for the trees” is an idiom to reminds us that we may not see the larger picture if we focus too hard on the parts. However, vertical SaaS may require you to reverse that idiom because it’s critical to drill down into each domain.
“Horizontal SaaS is seen as infrastructure software to which you add domain knowledge. But the expectation from a vertical SaaS company is that its software codifies a lot of domain knowledge,” says Hariharan.
CommerceIQ’s path to C20 in its first 15 months had a number of missteps which drove home the point that retail ecommerce management (REM), the vertical SaaS category it was creating, varied from domain to domain. “We went and sold our product to apparel companies and quickly figured out the product didn’t work there yet. We pulled back very quickly,” says Hariharan.
You had to work in every brand’s domain long enough to be able to codify the domain knowledge and requirements. “So if you don’t have TV customers, then you don’t know how TVs sell. Third-party seller issues are important to them; replenishability is not important for them. Understanding those nuances and baking it into the product becomes critical.”
Lesson 3 / Don’t mess up your product with hypergrowth
That makes it imperative not to be influenced by VCs on the board to aim for hypergrowth. “It’s doable in horizontal SaaS by throwing fuel (salespeople) on one solid PMF,” says Hariharan.
“You cannot do that in a vertical market because you have to grow the number of modules by building domain knowledge. You cannot buy your way into it”
“Two years ago, when all VCs were harping on growth rate, we said, ‘You know what? We have a path to doubling our revenue but don’t push us to triple, because that’s just the wrong thing to do.”
Unit economics for efficiency along with growth is twice as important in vertical SaaS. “Now, of course, in the last six months, people are asking for efficiency metrics, but frankly, this is what I showed in my series D fundraise in January 2022,” says Hariharan. “We had a burn multiple of just 0.62 in 2021 when everybody was going after growth. They were saying, ‘Why the hell are you burning only 0.62? Go burn at least 1 (the total net revenue)’.”
If you get into the common investor fallacy of pouring dollars for growth, you can mess up your vertical SaaS product.
“I actually did that in my earlier startup when I signed up a very large retailer and became a slave to them. We had to show triple growth and didn’t have true PMF,” recalls Hariharan. “So you end up becoming a professional services company masquerading as a product company. Then it starts to eat away at the company.”
His next startup, CommerceIQ, could never have gotten the trajectory it achieved if its PMF foundations had not been laid by setting C20 (20 successful customers) as the formative target.