My startup is on the early side of early-stage. How should we be thinking about growth?
“There’s a saying that there are only two phases in a company’s life, pre-product-market fit and post-product-market fit, and I’m a firm believer in that,” says Rothenberg. “If you’re pre-product-market fit then, seriously, you don’t want to be focusing on growth. A huge mistake people make all the time is trying to find a magic bullet growth solution that will prove they’re on the right track, but that doesn’t exist. Or if it does, it’s a spam tactic that won’t work for very long and will only do more harm in the long-run.”
Resources shouldn’t be dedicated to growth if your product isn’t on a clear path to sustainable user engagement and value creation. Everything you have and everything you do should be about finding that fit first, making sure there’s a real tangible need for your solution, reaching out and really talking to your initial customers to iterate fast.
“Growth expert Sean Ellis (CEO of GrowthHackers) says you should survey your user base before doing growth work. At least 40% of your users should say they would be very disappointed if your service or product went away overnight,” Rothenberg says. “If they don’t, then you likely don’t have solid product-market fit. Hitting that mark is a sign you’ve found your niche. And then it becomes all about pouring fuel — in the form of dollars or resources, of course — on growth tactics to hopefully return even more users and dollars.”
Importantly, the trigger metric here is engagement — a blend of healthy traffic, activity and retention. Startups are prone to taking only the first two as a green-light, and getting blind-sided when they discover they have a leaky bucket on their hands. This is why he recommends that early-stage growth and marketing teams focus the bulk of their efforts on customer acquisition, but regularly check in on the rest of their funnel. On a basic level this looks like:
Traffic → Sign ups → Key conversion event → Engagement → Retention → Evangelism
You can see why these initial responsibilities belong to marketing. Outside of core product experience, it’s difficult to meaningfully impact retention in a way that actually moves your business until you have a large enough customer base period, Rothenberg points out. Early on, most of your energy should go into awareness building and organic acquisition (press hits, brand-building, SEO, social, email) and testing paid acquisition channels. Retention is a little stickier to measure, but you need to have some sense of churn so you don’t commit to a faulty strategy. This becomes critical as the business scales, and revenue from past customer cohorts grows. If all signs are positive, then it’s time to sprint.
What can be shocking is how fast you need to shift into growth mode after this happens. Not only do you have to refocus on driving acquisition and conversion, you also have to rapidly scale the team responsible for it. Knowing that with the right product, this fast switch will occur, is half the battle. That way you can have a basic plan in place, ready to implement — a sort of “break in case of product-market fit” toolset behind glass.
This begs the question, what if you only have two or three people to work on this to start?
“You have to get ruthless about prioritization,” says Rothenberg. “Everyone needs to only be working on the small handful of things that will drive the bulk of the impact. The essence of strategy is deciding what to say ‘no’ to, and whoever is managing the team should help to identify those things, and provide air cover on everything that falls outside their reports’ critical initiatives.” Sometimes this will be the head of marketing, sometimes it will be a founder. Here are a few ways he’s learned to work smarter, not just harder in these ‘small team in a bunker’ situations:
Ask yourself, “What’s our business’ unfair advantage?” Identify and lean into to your unique strengths rather than trying to do everything.
“Doing what your competition is doing, just better or harder, rarely leads to winning a market. You have to find ways to do things differently.”
Build out a backlog of everything you think is worth testing and which key metric you expect it to move. Then rate each based on 1) expected effort, 2) expected impact, 3) probability of success, and 4) dependencies.
Here are some examples (using hypothetical data to demonstrate the method):
TEST: A/B test levers to improve conversion rate of signups for key conversion events.
METRIC: Sign up to purchase rate.
EXPECTED EFFORT: Low.
EXPECTED IMPACT: High.
TEST: Effectiveness of lookalike ads (ads targeted at new customer bases that look like your existing audience).
METRIC: Volume and ROI of signups.
EXPECTED EFFORT: Low.
EXPECTED IMPACT: Medium.
TEST: Retargeting ads designed to get people who visit your site to sign up.
METRIC: Volume and ROI of signups.
EXPECTED EFFORT: Medium.
EXPECTED IMPACT: Medium.
DEPENDENCIES: Design for ad creative.
TEST: A/B test adding social proof language to key landing pages to increase sign ups.
METRIC: Landing page visit to signup conversion rate.
EXPECTED EFFORT: High.
EXPECTED IMPACT: Low.
DEPENDENCIES: Engineering to implement, analytics to track
Spot patterns in your actions — inventory the low to medium-effort levers that have medium to high-impact and zero or few dependencies. (Not being dependent on engineering or other scarce resources within the organization to help you build experiments is a big one.)
Prioritize these levers first but don’t forget to tee up high-impact initiatives that do require working with other teams now. They take time to develop, and you can nudge the ball forward as you do other things.
Document your force-ranked list of priorities and tests and run it by your boss or best sounding board weekly. Be willing to move things around.
Always have a ‘water-line’ in mind — what you can realistically get done in the next week or month. Circulate this to the whole team. Don’t over-extend.
Working quickly on the things that matter most and are largely within your control to execute can get you pretty far, but soon you’ll need to think about growing your growth strategy. The other thing you can do is pick out the few people in your org who have been the most valuable for executing your current plan. Help them spread their habits to other members of their teams. Empower them to be evangelists for the way they work, helping to instill the growth mindset throughout the organization.
Okay, I’m ready to build out a team. What now?
First, the best growth teams are built at companies where the mandate and interest comes from top leadership. At Twitter, Dick Costolo was known for giving the growth team permission to rock the boat. At Eventbrite, Kevin Hartz was immediately open and supportive of building a cross-functional growth task force. The best structure is extremely dependent upon the type of business, and what works best in practice within each unique organization, largely contingent on the talent you have on board since growth talent is so hard to hire.
Second, it doesn’t have to be a big team.
At Eventbrite, growth efforts are spread throughout the org, with deep cross-functional work between product, marketing, engineering, analytics, and other groups. Rothenberg’s initial cross-functional strike team started with one person (him) from marketing, a product manager, a couple engineers, plus shared analytics, UX and design resources. This group was solely dedicated to user growth for a trial period — that’s how the team got buy in from executive leadership. It led to faster iteration and learning, and quickly became a fixture. The key, they found, was to empower a very narrow focus on not just broadly helping to grow the company, but on a specific growth goal.
Carve out resources that have a sole charter around growth, and know exactly which metrics they should be moving.
In Rothenberg’s case, this team was first committed to growing the supply side of Eventbrite’s business. A key initial mission was to test how to convert more event attendees into event organizers who would use the platform for ticketing. This came about through the team identifying one of their built-in ‘unfair advantages.’
Early on, Eventbrite gained momentum through the virality of event organizers inviting attendees to their events. After attendees bought tickets, these same people learned that they could use the platform to help organize their own events. This solved early customers’ pain points around their previously inefficient methods such as pen and paper, spreadsheets, collecting cash and checks, etc. More new event organizers invited more new event attendees, which triggered exponential growth.
With a core service that solved a genuine problem for organizers of all shapes and sizes, the growth team believed they could optimize this built-in viral lever to further drive near zero-cost customer acquisition. Through a multitude of tests and optimizations, these efforts ratcheted up conversions dramatically which helped to cement the case for a more permanent cross-functional growth team.
Much of what was tried, Rothenberg drew from prior experience.
At TaskRabbit in 2012, Rothenberg oversaw acquisition and retention marketing — but again, there was no formal growth team. Instead, he worked closely with product and engineering on ad hoc efforts to run experiments. In one instance, the acquisition marketing team partnered with engineering to build custom SEM landing page that dynamically updated page content based on the keyword used in searches. The idea was that this might improve conversion. So, for example, if someone searched “house cleaner,” they’d presumably load a page that would serve up potential house cleaners — as if by magic.
“We tested our way into this, starting with one landing page, which we then split out into four more tailored versions, each focused on TaskRabbit’s most popular service categories,” says Rothenberg. As you might expect, the tailored versions of the landing page converted better.
Seeing these results, they then split those four landing pages into even more tailored versions for subcategories. Again, conversion upticked. So they leaned in hard. Investing even more in their work with engineering, they built out granular smart landing pages that would detect nuanced ideas like “spring house cleaning NYC.” This would also automatically update key content on the page to make the experience more relevant for every visitor.
“This helped us improve conversion through the long-tail of relevant searches, ultimately driving our new customer acquisition,” Rothenberg says. Working together, marketing and engineering had produced better business results than either function had working independently.
Growth is best driven through tight alignment and coordination between marketing, product, engineering and analytics.
This was a tweak and definite improvement upon what he’d learned at his own startup SkillSlate in its early days. Striving to build the first local marketplace for independent service providers (think nannies, cleaners, plumbers), they knew there was an incredible long-tail of one to two person businesses out there that basically drove all their business through word of mouth. Most of them didn’t even have websites. But to succeed, SkillSlate needed to build a credible marketplace that would balance supply (these service professionals) and demand.
The team decided to focus on supply first, and they knew they needed to take a novel approach in order to have any chance at all. Interestingly, the founders had the makings of a nascent growth team, with expertise in product, engineering, analytics and marketing accounted for. They just needed to figure out how to move in unison.
Their first idea was to solve a critical need for service providers — namely, their discoverability. But that was going to be expensive, when really they needed to onboard supply at next to no cost (after all, they had no paying customers yet, and attracting them would take time).
What they did was build a web crawler that aggregated thousands of existing classified ads in one database. Then they scraped them for identifying info like phone numbers and email addresses. All the ads with the same data on them were lumped together and labeled the same business, allowing SkillSlate to use all that language together to describe what the person did. With the help of Mechanical Turk, they were able to turn these unstructured ads into much more cogent business listing with extra info like neighborhoods served, hours of operation, etc.
By publishing these listings to the internet, they essentially created the first website or web presence for many of the service providers they wanted to recruit into the marketplace. More importantly, it made their info sticky on Google, giving them better placement in search results, and driving better business.
“After we had all this, we were able to reach out to service providers to ‘claim’ these business listings in order to edit what we had put together and append customer testimonials,” says Rothenberg. “Many of them even stumbled on their own listings through Google before we got to them. This led to us acquiring several thousand providers almost immediately after we launched.”
While this cross-functional — one might even say “hacky” approach — worked to drive initial growth, it came at a cost. The system took longer to build than anticipated, and the team over-indexed on growth and scale at the expense of devoting our limited resources to solving customer pain points. “We had a really smart way to get people into our funnel on the supply side, but we didn’t use our runway wisely enough in driving hard toward product-market fit,” says Rothenberg. “To be honest, we missed the mark. I remember meeting with one of our investors who suggested we were trying to scale our way out of the product’s lack of engagement. In hindsight, he was 100% right.”
SkillSlate landed a pretty plumb acquisition deal anyway when it joined TaskRabbit, but he continues to wonder how the business would have thrived had they answered the product-fit question before rising to the challenge of scaling.
Looking back, there are a few things Rothenberg learned from all of his experiences:
Don’t try to scale too soon. “When I was working on my own startup, we tried to scale before having true product-market fit. We tried to brute force our way through rather than focus on core user value and engagement; it was a huge error.”
Don’t run growth independently through a single traditional functional group like marketing or product alone. You need folks with different skill sets and available levers working together and dedicated to the cause.
Always align your organization with the few key metrics everyone should rally around. Each person’s individual goals should directly ladder-up to at least one of these metrics.
Realize that every business is different. There’s not going to be a one-size-fits-all formula. Tailor these tactics to your market, but always drive hard toward simplicity. It’s really easy to overcomplicate growth efforts. Start with straightforward targets. Explain experiments clearly so that everyone can understand them. When you report results, make sure they make sense to everyone and lead to specific, logical actions.
Cool. Things are happening. But how do I know if we’re making progress?
Reporting growth results can be tricky because not everyone will care about the same things, even if you share the same metrics. It also makes a huge difference if you’re a B2B or B2C company.
For instance, if you’re B2B-focused, “The big idea is to provide real-time data on lead lifecycle stages and be able to track that efficiently,” says Rothenberg. “If nothing else, you want to report on volume and conversion rates between each stage of your funnel.”
Distill and circulate how close you are to meeting your monthly, quarterly and annual goals for these numbers and conversion rates.
“These metrics are important because they provide you with insight into your strengths and shortcomings so you can pivot your marketing plans,” he says. “You also want to track these conversions on a campaign basis so you can see what types of campaigns are the best for generating and/or converting different types of leads.” Some terminology is helpful here:
Leads: Prospects who express very basic interest, such as signing up for your product or filling out a lead form on a piece of content.
Marketing Qualified Leads (MQLs): Prospects who express much deeper interest, metaphorically ‘raising a hand’ to be contacted. This is often based on the quantity and types of content they have consumed, or actions they have taken such as watching a recorded product demo and/or viewing your pricing page.
Sales Qualified Leads (SQLs): Prospects who have been reached out to, researched and vetted by your sales team.
Opportunities: Prospects that are in active discussions with your sales team and are one step away from becoming customers. (You want to first convert your MQLs and SQLs into opportunities, for instance).
All four should be monitored in any B2B growth plan. Let’s say you’re 75% to your MQL goal, 10% to your SQL goal and 50% to your opportunity goal. You’ll know from just looking at these numbers that you likely need to spend your time pumping up your MQL to SQL conversion, or figuring out where your higher-value opportunities are coming from. You can then refer to your campaign reporting to decide which type of programs to run that will help you get there.
When it comes to B2C businesses, what works tends to vary more. To illustrate some tactics, Rothenberg cites an example of marketplace businesses.
Marketplaces tend to have a North Star metric of Gross Merchandise Value (GMV) — the value of all goods and services transacted on the platform. This metric has several inputs and funnels:
Supply side: Traffic → Number of registered sellers → Number of goods listed → Number of goods sold * Avg selling price = GMV
Demand side: Traffic → Number of buyers → Number of items bought * Avg selling price = GMV
To get to transactional revenue (the largest source of revenue for a marketplace), you typically multiply your GMV by your take rate (your service fee).
So, for a marketplace, the high-level things you’d care about are: how much supply and demand you have (within geographic areas like cities, or within categories like books, electronics, etc.). This leads you to care deeply about how many of your items sell and how quickly, and how efficiently you’re matching supply and demand (or your liquidity).
Once you have a basic equation for both the supply and demand sides of the marketplace, you should map your metrics for each. You’ll likely identify areas that are more or less healthy than others.
From a high level, it may seem like you have enough supply, but you need to drive more demand. If you go a level deeper, the reality could be that you have twice the supply you need in a given category, but half of what you need in another. Identifying and balancing these situations will help you drive liquidity. Going deeper still, you might see that you have healthy traffic and signups, but your conversion to goods listed is low. This can help you know where to invest more time, money and effort to drive everything else up.
The progress you’re making on your plan needs to be monitored constantly.
“Campaign reports are best if updated weekly and distributed to the internal marketing team,” says Rothenberg, who recommends adding gauge charts to your CRM to get at-a-glance insights. “This will allow you to make mini-pivots in your strategy backed up by data.”
On a monthly basis, compile the progress you’ve made against your plan, key learnings and key challenges, and share that report with sales, he says. Briefly bullet out the campaigns you plan to run over the next month based on these lessons. This will keep communication healthy, and hold you accountable to constantly improve.
It’s easy to get too granular in your reporting — which wastes time. For example, open and clickthrough rates on emails can be helpful for the right people on the marketing team to optimize what they’re doing. But they’re not helpful for broader funnel analysis, Rothenberg says. They don’t need to be passed along to other teams.
“Everyone who needs to know real-time data should have access to the right dashboards in your CRM,” he says. “But we don’t send out too many reports to people via email. We make a point of covering the key ones in weekly, monthly and quarterly meetings so we can provide the necessary commentary and break down complex topics in real time. We don’t want to just show people numbers, we want to describe what the numbers are telling us and discuss what we’re doing about it.”
Keep in mind that setting the wrong goals can be worse than setting no goals at all. Sure, this runs counter to the analytical growth approach, but wrong goals can really do some damage. People’s nature is to drive toward targets. But if these targets aren’t aligned with what truly drives a desired outcome, then everything can break apart.
Early on, before you know what drives your business with great precision, it’s very helpful to be flexible on your targets and to regularly adjust and reforecast at regular intervals along the way. “This is an ongoing process for us at Eventbrite,” says Rothenberg. “We’re always gaining new understanding of our business levers and how we can work smarter toward what will actually drive results for our customers first, and then our business.”
What tools do you use to do your funnel analysis? And how do you act on the results?
“When working with early-stage companies, I typically use some combination of third-party tools like Google Analytics for traffic analysis, Mixpanel to visualize funnels and Optimizely for A/B tests,” Rothenberg says. “These tools tend to work well when the business is ramping up, though during scale stage most companies begin to rely more heavily on internally-built tools, such as a proprietary A/B testing platform, which are made possible as the business scales and its needs change.”
You also want to spend time calculating the unit economics of your business. What is the value of each user? How is that trending over time? How much should you be spending to scale users?
It’s vital that you understand the relationship between customer acquisition costs (CAC) and customer lifetime value (CLV or LTV for lifetime value) for your business. This is how you’ll ensure that user acquisition is profitable.
“At the earliest stage of a startup, you’re just trying to acquire enough users to get to product market fit and see if the unit economics are going to work,” says Rothenberg. “You want to keep it relatively simple at the stage, including direct spend and things like promotions.”
The goal is to be directionally in-bounds when spend is low to figure out which channels will yield profitable acquisition.
At scale, CLV should include almost everything: headcount, direct marketing spend, branding spend, discounts and promotions, and more. Take into account company overhead (your lease, support functions, cost to service recurring revenue, etc in order to get to your true gross margin LTV. The simple equation is:
CLV = (Revenue per User * Gross Margin %) / Customer Churn Rate
You really want to dig deep here to make sure what you’re doing will pay off long-term.
“A common mistake that I see startups making is that they calculate CLV once and think it’s set in stone,” says Rothenberg. “As acquisition scales outside of early adopters, nearly all businesses eventually acquire lower-value users. This degrades CLV and can lead to an upside-down situation where CLV ends up being lower than CAC.”
Make sure to update your CLV estimates at a regular cadence, quarterly or bi-annually to start, to make sure you’re not spending more than you’re getting from each incremental user. “It sounds obvious, but many companies fall prey to this, and scaling negative unit economics is a quick path in the wrong direction,” he says.
Keep peeling the onion to understand the unique drivers and metrics of your business. Track those obsessively.
“Conversion is a lever to push on every day,” he says. “Over my career, I have run hundreds of A/B tests on key landing pages and user flows in order to continually iterate and understand what improves conversion all the way from signup to the bottom of the funnel. As you scale, your highest leverage tends to be deeper in the funnel; I have repeatedly seen how some of the biggest, most overlooked opportunities are deeper in the funnel.”
For example, at TaskRabbit, his team ran a number of tests throughout their funnel and found that the biggest impact came from improving the match rate between people posting tasks and the taskers actually doing them. With this hypothesis, they made product changes to increase the number of successful connections, and the results weren’t surprising. This ultimately led to the company’s pivot away from customer bidding to on-demand service, which quadrupled the company’s revenues.
“Credit goes to the team there after I left, but it’s a prime example of a business figuring out what really matters, running experiments in that area, and then quickly incorporating those learnings into its core product,” says Rothenberg. “The business taking off after those changes really spoke for itself.”
As a business scales, it becomes increasingly vital to not only focus on filling the top of the funnel and moving customers through, but on how you engage and retain those customers. “The larger the percentage of your revenue coming from returning customers, the bigger the opportunity becomes to positively impact your overall revenue through modest gains in your returning customer base,” says Rothenberg.
This can be a very difficult lever to move, but it starts with understanding why your most valuable customers stay with you, and conversely, why some of them leave.
You have to nail that shift from focusing on sheer growth to full engagement. Making it too late is a costly mistake.