Contact

Privacy policy

Contact

Privacy policy

Aha moments: MHV's Peng on retention and viral coefficient for startups

Sep 6, 2023

Community notes by: Sina, CTO/CEO of Learning Loop

Disclaimer: This post is not written by AI. A Learning Loop member wrote them while listening to the talk. Therefore the notes aren’t direct quotes. The speaker isn’t responsible for how these notes or the tone are perceived/interpreted. If you have questions about specific parts of the post, email team@learningloop.org for clarification

Conversation between:

Came to the event primarily curious about this:

Peng has founded match.com that shaped how people date, Interwoven (CMS) before CMS was mainstream and took it public in NASDAQ, and Ecentuate (identity management) and sold to IBM. And now running a VC firm for the past 10 years that a lot of founders I know respect a lot. Sure he has had teams that have worked with him in all of this, but for a founder to found 4 great things in a row is too statistically significant to be a coincidence. He must have superior mental models. How does he look at the world?

Notes:

  • How do you view the world? 

    • You have to observe ridiculous states of the world and do something about it

    • A state of the world that makes you go “what? That makes no sense”

    • Going thru printed classified ads back in the day and marking it with a pencil to call the person later vs digitising it became match.com

    • Thousands of people trying to produce stuff on 1 stuff and constantly taking the website down as a consequence. That’s ridiculous, you need a CMS.

    • When he was younger, he looked at the ridiculous things in the world in narrow areas. 

    • Now he looks at larger scale. e.g. 1/3 of kids alive today won’t be literate when they grow up. Nobody is working on this although it’s crazy.

  • Where did R+K>1 come from?

    • reference: https://www.monkshill.com/content-hub/series-a-startups-take-note-r-k-1-is-your-formula-for-success 

    • LTV/CAC was leading some of the portfolio companies to death, because CAC is controlled by Google/FB/tiktok etc. and it’s out of your control. 

    • Five years ago we started thinking in terms of relate retention and engagement and how those things need to behave for business to make sense

    • Objectively how much do the client customers/client/employees love the business, and how many new clients/customers are they bringing 

    • If R+K < 1 then the total number of clients/customers keep dropping until you die

  • What’s the minimum sample size and timeline for measuring this?

    • Statistically at least 100 users and at least 6 months for you to know As a VC I prefer 12 months, and of course I prefer a sample size of 5000 clients/customers than 100

  • Many b2b and enterprise focused companies optimise for revenue and probably not vitality/engagement/retention in the b2c sense. Does R+K>1 still apply?

    • He’s a b2b first guy and most businesses he’s built and invested in are enterprise and b2b

    • They initially thought this model doesn’t work for b2b

    • He still doesn’t know if it will

    • But he has CEOs that are thinking in terms of B2E2E2B. The E is employees

    • Creating and measuring vitality in terms of the number of new clients the employees in your target organisation bring onboard 

    • e.g. Slack sold to 3 people in a company and they brought everyone else in that org onboard, and eventually other orgs 

    • A network within the company

    • Another angle is community building

  • Audience question: how do you think about retention in the case of platforms like Carsome where user may buy a car from them every 5 years?

    • By creating a a new type of relationship beyond that transaction

    • Engaging customer via community or outbound messages that helps the business learn more about the customers and their needs and adding new value to customers that the customers may not be asking for but would appreciate/pay for

    • It’s like everyone having an agent

  • Audience question: at what point do you decide if a company is worth investing in, since building a community and network effects isn’t easy and require lots of resources.

    • In the end it’s about the efficiency of the business

    • If you have a highly profitable without network effects he will still consider to invest, but then the question is can you scale? 

    • Can you scale just as profitably 

    • Can you scale without a community

    • Your cost of sales goes up independently of you when you don’t have network effects, just because of market forces

    • You lose control over your business model

    • We don’t like business where the founders dont have control over their business 

  • Advice for early stage companies with little resources? 

    • There’s a whole spectrum

    • If you’re b2b, your LTV/CAC is probably high enough, and you can take your time to build a community. And also you need to engage your employees a lot before you can build a community.

    • But if you’re building something that needs a community in order to improve your unit economics, you need to build it upfront. you can’t build a business first and then figure out the R+K>1, the community becomes a prerequisite.

    • R+K>1 kicks in right after pmf, when you need to “grow” the business. You must derisk PMF first.

    • R is a great product. K is also a great product. 

    • If you build a great product, people use it, stay, and bring other people to it.

    • Build something that’s as retentive as possible and as viral as possible.

    • You may have a product that can be viral and retentive but you haven’t built in structures that leverage it. e.g. lock in factor, how many steps does it take to invite someone? 20 clicks? Vs 2

  • How do you balance paid growth vs vitality?

    • There is a whole spectrum from free new members vs full CAC members

    • When you have 0 users, you can invite all your friends. A very slow way to crank up a system. You pay money to get people on the system to see if your R and K are any good.

    • It’s a learning process to use R and K to jumpstart your engine and see what’s equation producing. If you get users and R+K<1 then there’s no point getting more members.

    • If there’s a leaky bucket, dont pour more resources into it.

    • Use your resources to optimise your R+K>1 to get the product model right

    • You wanna get to R+K>1 while your LTV/CAC is at least 2 (LTV/CAC may drop to less than 2 at some points in your business’s lifetime and if R+K<1 or = then you don’t have a business anymore)

    • There are semi paid for things like “buy 2 for 1” and etc. that are okay. You can use them but it’s best if it’s viral by design (free new members vs full CAC members)

    • The best ones are people telling other people because it’s good

    • Whatsapp and Instagram are value-less if you use them by yourself. Viral by design.

  • Audience question: implications for Web 3.0 since it’s super community based?

    • He doesn’t know much about it. But the good thing is that you can see how users transact and use those linkages to design viral behaviours.

  • Audience question: how do you prioritise product, customer support, servicing, etc.?

    • “I’m old school, all of those are product”\

  • Audience question: how do dating companies do R+K>1?

    • Dating is not a great example because if you succeed your R+K drops

    • Apps like tinder are more profitable because they’re not out there to succeed in matchmaking, or at least not for longer than a few days

  • What are some odd examples of companies that looked like they couldn’t ever have a community. Do you know any companies that looked like they couldn’t have a community but managed to build a great community

    • semiconductor, supply chain optimisation

    • People who care deeply about esoteric topic. Enough of them can form a community

    • People collect all sorts of stuff

    • Community is about finding commonalities 

  • Audience question: in B2B context, you have different types of users. The people who setup the system, the regular employees who use the system, the decision makers. How do you figure out community?

    • It depends because if the product is counting days off, then there’s probably not much room for community building 

    • But if there’s some element of learning involved, then you can certainly build community around that

  • Audience question: are there any indications for when your current business is ready enough for virality?

    • Your retention is as high as you can make it. Before the retention is high, there is no point pushing the vitality. 

    • You don wanna get vitality that doesn’t lead to retention

  • Is there a benchmark for R in SaaS?

    • Depends on the business. Some businesses have much higher R’s, some have 0.5 and that’s good enough 

    • In SaaS, he recommends companies to charge for shorter periods of time to learn if their retention is good, unless they have an easy way to look at usage. For example if you charge for the year, and assume paying for 12 months means 12 month retention, you’re measuring the wrong thing

  • Audience question: what’s a good indicator that a company has found PMF?

    • Easier than you might realise

    • The market almost pulls your product

    • Once PMF happens, you realise you don’t struggle to sell the product

    • If you a small ARPU business (which means you need a large volume of sales going), you should see sales just happening

    • If you’re still struggling, it means you dont have PMF

  • Audience follow up question: could the struggle just mean you’re in a competitive market?

    • If every time you lose a sale is because of a competitor, then you’re in a market too busy for startups

    • Other than that, if you’re struggling to sell and when you ask why people give you different answer, you’re probably not looking at a ridiculous problem in the world, maybe just an interesting problem to solve

  • Audience follow-up question: is it just gut feel then?

    • Nope you can look at data

    • Here are 100 people who are qualified. what % said no and why. If all of them say “too expensive”, that means you have a pricing issue. But if they say they dont need it, then you objectively dont have pmf in that segment.

    • Every single yes or no, you must know why. 

  • Audience question: what creative ways have you seen people foster a community model?

    • I’m not an expert on community building but I found user generated video answers to FAQs pretty interesting. Not just testimonials.

    • Anything a new customer asks should be answered by several past customers and publicly searchable/discoverable 

    • Incentivise this. Customer needs a discount, give them a discount but ask them to answer a question via a video etc.

    • When you film a testimonial, dont just let the customer talk, you ask them questions. Usually you ask them a bunch of questions

    • Include a few in the FAQs

  • Audience question: how do you think about innovating in emerging markets that are using old tools?

    • Find out what they’re doing now, do exactly that but digitised to derisk PMF asap, then build the more ambitious, opinionated stuff that you wanna

  • Audience question: What do you mean by derisk PMF?

    • If you’re going on a mission in a jungle and there’s a turn that has a 80% chance of killing you, your job as a leader is to ask “how do I not take that turn, how do I not get killed?”

    • In startups, that’s PMF risk.

  • When a leader sees that risk of death, has choices, but ignores it and moves towards it, what do you think of that?

  • Audience question: As a VC, to what extent do you try to validate a company’s R+K>1 before investing?

    • This is not religion at our VC firm. 

    • We’re all former CEOs, we don’t treat each other as kids. 

    • Any company that is going after a mass number of users, I’m looking at their retention and vitality. 

    • If they’re not there, and worse, if the founders have no clue they should measure that, I’m not interested. 

    • Surprisingly many founders don’t measure R and K. It’s like flying a plain and not knowing how the lift works.


References

Sina's recommendation (related to the points discusses)

  • The cold start problem (read the chapter on how slack got employees in target companies to bring other users onboard) https://a16z.com/books/the-cold-start-problem/

  • The embedded entrepreneur book (on building audience-driven businesses) https://www.embeddedentrepreneur.com/