So far, we have built up a marketplace and we have seeded supply and demand. Now we need to scale it up.
Longish post. Intended as a comprehensive resource to be referred to as needed.
One - How do we know we are ready to scale up?
We have very healthy numbers in our business.
We have some strong new geographies / whitespaces to open up and there is enough empirical evidence which indicates that our hypothesis is solid.
Serious competitors are emerging.
Two - Going back to Constraints
Deciding to scale-up is a good time to see what limits an incremental transaction on our Platform? Think of it as a see saw. We will not grow if we are adding weight on the Heavier Side. It’s a classic resource allocation problem.
Whether it’s supply or demand is a harder question to answer. As an example, Uber/Ola will always be supply constrained because the harder piece of the equation to build is the vehicle. AirBnB will always be supply constrained because again, the harder piece of the equation is the house. On the other hand, an UrbanClap is almost always demand constrained. In the middle of these extremes, we have a Swiggy, a Zomato. They are by and large demand constrained but for some cohorts, they are supply constrained. Things can become confusing quickly.
However, most often, one finds that marketplaces are supply constrained. Far too often, RoI on incremental supply acquisition will outperform RoI on incremental demand. It’s a question of optionality for end customers. It’s also a great test to figure out where the constraint lies. There may be other empirical tests / heuristics to figure this out but they may be unique to each situation. An example will be correlation of NPS with incremental demand or supply side growth. If you are getting better NPS by investing on demand - that answers the question succinctly.
Three - Scaling Levers
Now, we enter into the territory of handing money to Googles and Facebooks of the world to get the word out to our customers1. Lets look at the levers:
PR: Interesting beast. Press, bloggers, influencers etc picking us up will generate a lot of free publicity. A sane PR budget is usually a good thing. The issue with PR is that while it can show huge potential - empirically its always been (in my experience) a good to have - not a need to have. Usually it serves as a great complement to the other levers we discuss below.
SEO: SEO is usually a source of free traffic. It is important to look at this as the lowest hanging fruit in terms of bang for buck. Sure, performance marketing looks attractive (and it will be a large part of what works most likely) but if we have a decent PR and content game, this is a no-brainer. It’s not a variable cost, typically pays for itself within months and from whatever data I have seen, leads to higher customer lifetimes. YMMV2.
Better Conversion: This is also a relatively low(er) hanging fruit. If we have already spent on driving traffic to our marketplace, we are indeed doing ourselves a huge disservice if we aren’t converting that traffic effectively. Removing friction and maintaining engagement are two key drivers here. However, this is a caveated lever. This is like a computer - GIGO3 applies. Unfortunately, conversion optimization is almost always a complement to another strategy (like PR above). By itself, it is not something that can underpin a scaling-up exercise.
Direct Sales: Direct sales is relevant for a few kinds of businesses (think of Zomato signing up restaurants). Having more FoS/outbound sales aids with evangelization. It also potentially sets you up for a sale down the line even if a prospect didn’t purchase today.
Referrals: Referrals are a function of Word of Mouth and a well-engineered referral program can be a great source of user base growth. However - the trouble with referrals is that we will likely see much higher drop-offs after the initial transaction. I still don’t think there is a perfect referral system out there. However, it can be very, very powerful on the supply side.
New Markets / Whitespaces / Geographies: Many marketplaces have a much easier structure, especially when they are enabling local / hyperlocal commerce. Simply finding a new area to open up and expand to is a valid scaling strategy. I love such businesses. They are what I call “business in a box” models. Here is the playbook. Rinse, repeat. Unit economics are easier to figure out, failure standards / checkpoints for intervention can be easily defined, course corrections can be made cheaply. And if it’s not working out - we can withdraw. Critical skills needed of us here are a deep knowledge of the market, deep anticipation and quick action on sources of operational leverage; and an ability to look at things dispassionately4. Again, UrbanClap is a great example of this. City by city, careful execution, no pell-mell assaults. Eventual end state is an acceleration in achievements of milestones in each new market because our presence in places where we operate is building latent demand in places where we aren’t.
Performance Marketing: Like the baniya that I am, I have saved the most expensive way to go for the last5. As we scale and refine our parameters for performance marketing, our funnel will look more and more weighed towards performance marketing. So we may go from mostly referrals/WoM to mostly paid marketing. When done properly, it’s a great force multiplier. When done poorly, it is a sureshot way of spending yourself into oblivion6. There are geniuses in this field. However, to our intrepid entrepreneur, I will offer only one limited input here - Don’t think of LTV when deciding on these investments. Think in terms of contribution margin. Trying to measure success of performance marketing baselined on LTV is likely going to lead to overspending and over-estimation of value (given that LTV itself changes and is uncertain). Evaluating success of PM based on contribution is much more rooted in real world cash flows. I would argue that it is a more sensible way to look at success / failure.
Four - Maintaining Quality
I will break this down into levers but I think this is mostly common sense.
I would caution us against letting pressures of growth make us think that existing users (supply and demand, both) are in the bag. They are not. If they are capable of backing us early on, they can just as easily vote with their credit cards and time and move their business elsewhere.
These are the guys who are supporting our marketplace. This cohort of users are early believers in our offering and will have a disproportionate voice. We don’t need to bend over backwards, but we need to keep them from feeling neglected and we need to make sure that our quality is not suffering in our mad rush to grow.
The levers below apply mostly on the Supply side. There are a few that should apply to the demand side as well, but that usually is only in cases of demonstrated fraud and abuse of the system.
Define / search for indicators of quality - The most successful marketplaces are able to identify a few signals that predict long term quality and then actively help the supply that fits these indicators succeed. I read somewhere7 that TaskRabbit gave its best Taskers a Meyers-Briggs test and were able to see which personality types were most successful on the Platform. Then they began to double down on these types. Think of the intellectual leap needed to make this happen even though it is fairly obvious when we read about it now.
Maintaining training / on-boarding standards - make sure we train the supply so that they are aware of whats expected of them. It’s expensive but pays for itself in the long run.
Product Standards and Incentives - publish clear standards and rules to be adhered to. Incentivize on out-performance, penalize on underperformance. Understanding that quality itself is a differentiator is really important here.
Reviews and Ratings - what we engineers call a self-sustaining system. Helps build trust and maintain quality. If you are getting too many poor reviews / ratings, perhaps a conversation is in order.
Ranking to promote the Best Supply - one of the core propositions of a marketplace is discovery. It is important that there are enough filters that allow the customers to rank the discovery by the parameters that matter to them. You can also seed the supply in the search results in an intelligent manner. The place where I see this most often is Urban Clap. They show me the service providers I have used earlier, and then seed the new supply quite intelligently. I almost never see the same names in the results 4-7 on any search. It is a rather smart way of maximising LTV.
Lastly - Invest in Customer Service. Mistakes happen. That someone will screw up royally is not a question of if but when. And when mistakes do happen, it is critical to make sure that the bad experience is made up yesterday.
A great example of how to NOT do things is Zomato. If you receive bad or cold food from Zomato, they almost never take any ownership. Everything is the restaurant’s fault, never Zomato’s. And their first line of Customer service is not to offer a refund but to offer an opportunity to write bad feedback on restaurant page. Only when you fight or escalate do you get a fair shake. There is a reason why OpenTable is a gold standard, and irrespective of people waxing eloquent about how Zomato reflects a mature ecosystem in India, restaurants were agitating against Zomato.
Maintaining quality is a long, ongoing, never-ending process. There will be unforeseen detours and constantly shifting expectations from the ecosystem. EVERY successful marketplace has a full-time team focused on maintaining quality and Customer experience. Scaling comes with its unique challenges but if the experience on either side of the marketplace is compromised, we will end up ceding hard fought ground to our competitors.
Thoughts and feedback welcome.
Unlike last time, I am not bothering with images this time. It took too much time and I found most online tools to be really a bit unusable for my purposes.
Your mileage may vary.
Garbage in, Garbage Out.
i.e. Knowing when the current playbook is not working.
In some ways, this is the easiest lever to use because it generally involves throwing a sack of VC cash at the problem. My concerns lie around durability and maintaining cash flows but then I am obviously a member of an endangered species.
Remember housing.com?
Perils of voracious reading. I will update with the source once I find it.