Notes for a startup: 6 questions to help you pick the right investor

This is from a great article on Medium, called 6 simple questions a venture capitalist should ask before making an investment decision.

I've taken the article and thought about what it means for which investors you should pitch.

A man called Par Jorgen Parson, who's invested in companies like Spotify, shared a bit of analysis he conducted into his own investment decisions and their success. 

He found SIX questions hold the key to his success, and so I've turned this into the SIX questions an entrepreneur should ask before pitching an investor:

  1. Has she invested before in a company addressing this specific customer category?
    eg, Consumer to Consumer, B2C, Prosumers, SMEs, Large Corporates, Government, Defense and similar.
  2. Has she invested before in a company addressing the same industry vertical?
    eg, Online Media, Adtech, Consumer Fintech, E-commerce, Security Software are examples of industry verticals that are specific enough.
  3. Has she invested before in a company at this stage of development?
    eg, Seed, A-round, B-round, C-round, bootstrapped mature company taking its first external capital.
  4. Has she invested before in this specific geography?
    eg, Stockholm, Berlin, New York, rather than Nordics, Germany, East Coast.
  5. Has she invested in this Founder team or members of this Founder team before?
  6. Has she co-invested before with the other VCs in the deal?

Here's the interesting bit: 

  • as long as Parson said YES to 4 or 5 of the questions, it was a good investment.
  • If he answered YES to less than four of the questions "I should not have done it"
  • If I he answered YES to all six questions, it didn't improve his investment outcome. Instead it "dulled it to just below what I would have generated without considering the questions at all." 
  • Other questions that turned out to have zero correlation to outcome, eg, “Have I invested in this business model before?” or “Is this my typical round size?”
  • If he had followed this rule from the start, he would have made a "staggering 50% improvement from an already decent performance base!"
  • Note: sample size is just over 30 companies so this is scientific, just anecdotal.
  • Note that investments in Parson's portfolio like Spotify, Avito, iZettle, Widespace and Pricerunner all fit this model.

Parson tries to explain all this:

  1. Minimise your unknown unknowns
  2. Don't just invest as you have before: "if you are always inside the box, how could you then realistically think outside of it?"
  3. "if you want to take a very high strategic and calculated risk in one dimension — like building your success on a new emerging model or distribution platform (think Google Adsense in 2003, Facebook in 2005 or Appstore in 2008) — then you may want to minimize risks in other dimensions. Outsize returns often come as a result of capturing such seismic shifts at the right time."

Final point from Parson: he says the key denominator of success is "entrepreneurial team quality and true grit. All other factors are, in my experience, deal specific". 

What this means: if you're approaching an investor, quality her/him by asking yourself and them these 6 questions. You'll save her/him and you a lot of wasted time and effort.

The platform revolution

  • We think the shift from pipes to platforms is the most important business shift so far this century for one very simple reason:

    • platform businesses disrupt and dominate vast industries within a few years of launch.

  • See our research on the Future of Business for more.

  • Like all posts at The Future Is Here, this post is a work in progress. It's a summary of the research, work and thinking we've done on platforms.

Its starting point is the work of Sangeet Paul Choudary, including articles and his books, Platform Scale and Platform Revolution — co-authored with the two MIT professors who made two-side markets famous in a HBR, Geoffrey Parker and Marshall Van Alstyne.

In case you prefer to watch than read, here's a keynote he gave about it.

In the 20th century, to get bigger, or "scale" as people like to say today, a business had to own the means of production. To get bigger, they had to scale internally. 

Now, in the 21st century, that's changed. The fastest growing companies — and most valuable — in the world scale externally.

A good way to see that is the difference between pipes and platforms

Instead of creating the supply, and selling it downstream to consumers. platforms don't own the means of production. Instead, they are like marketplaces. They provide the infrastructure, the space, the place for producers to sell to consumers. 

The pioneers, and now classics, of this platform model are: 

Airbnb — world's largest hotel company, owns no hotels

Alibaba — world's largest retailer, owns no stock

Uber — world's largest taxi company, owns no taxis

Rather than produce, these companies are platforms for other producers to sell to consumers.

New ways to create and capture value

All businesses are built to create and capture value. Platforms just do it in a new way. They're key raison d'etre is to make the exchange of value between producers and consumers are friction-free as possible.

Why is this happening today?

The key trends driving the rise of platforms are:

  1. We're all connected
  2. De-centralized production: it's easy for the small guys to be producers (and platforms mean they can compete with the big guys)
  3. the incredible power of artificial intelligence and algorithms — which makes matching possible
    (just think how hard a cab company's telephone operator woulda had to work to match up all those people wanting cabs with the cabs on Uber)

Core take-out 1: "interaction first"

Successful platform businesses are "interaction-first" businesses. The most important thing they do is make the interaction between the producer and consumer as frictionless as possible.

The core interaction is the crux of a platform. 

Your central organising principle should be: 

"all design decisions should ensure the repeatability and sustainability of the core interaction."

Your goal

The goal is to put an apple core at the top of this mountain.

The goal is to put an apple core at the top of this mountain.

The goal of the platform is to: "maximize the repeatability and efficiency of the core interaction"…

to optimize the flow of value and currency

Beware features and added functionality:

"The prioritization of features and functionality, community management, and marketing, should be based on their ability to make the core interaction more efficient and repeatable."

More on core

When you're designing the core interaction, think about the core value unit — which is the equivalent of the ride on Uber: which contains the key data to make the interaction happen: 

  1. supplier listing (note: screened, checked for quality)
  2. geography
  3. availability 

The platform's role is to match that with the consumer, whose data is similar (if opposite):

  • consumer listing (note: screened, checked for ability to pay)
  • geography
  • availability (or, in their case, desire, as in "I want a cab now")

When designing the platform, design the core interaction first, then lay out the others, which Choudary calls "edge interactions".

What is essential data for the core value unit? Who, what, where, when… and how good (and who says how good it is — included this because of the importance of quality/trust for consumers: how is this indicated? Consider Uber's, Airbnb's, and Amazon's ratings & reviews.

16 core questions for designing the core interaction

A brief respite from the half-eaten apple images.

A brief respite from the half-eaten apple images.

  1. What is the core interaction that your platform enables?
  2. What is the unit of supply created on the platform? What is the supply or inventory created on your platform?
  3. Who are the producers of value? What motivates them to produce?
  4. Who are the consumers of value? What motivates them to consume?
  5. What channels are used by producers to create value on the platform?
  6. How does the platform manage access control for producers on these channels?
  7. What channels are used by consumers to consume on the platform?
  8. What filters?
  9. What tools and services should you provide to enable the core interaction?
  10. What tools and services should you provide to enable creation?
  11. What tools and services should you provide to enable curation?
  12. What tools and services should you provide to enable customisation?
  13. What tools and services should you provide to enable consumers to consume?
  14. How do the tools help your platform to pull, facilitate, match?
  15. What currency does the consumer provide to the producer in exchange for value?
  16. How does the platform capture some portion of his currency?

9 core questions for designing the core value unit

  1. What is the core value unit? What is the unit of supply on your platform that defines value for a user?
  2. Who is the producer?
  3. How does the producer create the core value unit?
  4. Who is the consumer?
  5. How does the consumer consume the unit?
  6. How is the quality of the unit determined?
  7. What is the filter used to serve the unit to the customer?
  8. What consumer actions help create the filter?
  9. How does the consumer consume the relevant units?

Core take-out 2: platforms are two-faced

"Two-sided markets require two companies, often with completely different challenges" 

For a platform to succeed, it has to do 3 things: Pull, Facilitate, and Match.

  • on the producer side, Pull: the challenge is to encourage/incentivize producers to regularly produce core value units, each of which must have the right dat so that the system can match that with the most relevant customer.NB: units that cannot be matched do not contribute to activity on the platform, and the platform only has value when it creates value… which is when there is activity. 
  • on the consumer side, Match: the challenge is to scale quantity and quality

To really succeed it should also Facilitate as many aspects of the interaction as possible, exchanging value between producer and consumer. Consider: 

Uber facilitates value exchange: 

  1. transfer of information (a form of value) from producer to consumer
  2. transfer of transport-as-a-service (a form of value) from producer to consumer
  3. transfer of money (a form of value) from consumer to producer
  4. transfer of rating (a form of value) from consumer to producer, and from from producer to consumer
    This creates "persistent value", and creates a feedback loop.

NB: monetization is dictated by which transfers are captured and tracked by the platform.

NB: if your platform owns the end-to-end interaction, you're more likely to capture the user's input on quality — and hence scale quality as well as quantity.

NB: Interaction ownership is critical to create a sustainable platform business. (If possible, prevent off-platform collusion, through carrots mor than sticks.)

Rules, Tools, and Trust — the 3 keys to facilitate

The psychologist Jonathan Haidt thinks of the subconscious as an elephant, the conscious as the rider. The rider is mostly in charge. The elephant will follow her/his lead. Unless it wants to go somewhere else. The concept of "nudge" is about designing pathways to make it easy for the elephant to follow the path we want it to go down.

The psychologist Jonathan Haidt thinks of the subconscious as an elephant, the conscious as the rider. The rider is mostly in charge. The elephant will follow her/his lead. Unless it wants to go somewhere else. The concept of "nudge" is about designing pathways to make it easy for the elephant to follow the path we want it to go down.

To facilitate effectively, at scale, you don't need to referee and personally oversee every core interaction. (Doh!) Instead, you create a culture where the right sort of behaviours happen. You do that through:

  1. Rules
    1. Nudge users to compliance and the desired behaviour
    2. Create simple pathways for producers/consumers to follow
    3. Use cues, notification, feedback
    4. Core questions: 
      1. Who do we want to create core value units?
      2. Who can create core value units?
      3. How can they create core value units?
      4. Can we make that simpler, reduce steps?
      5. Can they do it passively?
      6. What differentiates a high quality unit from a low quality unit?
  2. Tools
    1. "Kill the skill barrier" with tools that make anyone look good
      think Instagram's filters: making crap photographers look like they have talent since 2010
    2. Allow for "emergence" — when users develop behaviours you weren't expecting
      Watch out for it too
  3. Trust
    1. This is quality control. 
    2. Have mechanisms that identify, differentiate and encourage good behaviour
    3. There are 7 ways to confirm trust online

The 7 Cs of trust

 

You can create trust in 7 ways

  1. Confirmed identity
  2. Centralized moderation — in the early days. This will give way to…
  3. Community feedback
  4. Codified behaviour — ie via implicit rules
  5. Culture — create a good one
  6. Completeness — think of LinkedIn's progress bar (could you do that for producers/consumers?)
  7. Cover — think of Airbnb's insurance

The 5 types of payment

There are 5 ways a consumer can "pay" a producer — that is, offer value: 

  1. money
  2. attention
  3. reputation
  4. influence
  5. data

A platform needs to make that value exchange as friction-free as possible: 

6 of the best ways to capture value on a platform

  1. % — take a discount/commission
  2. Charge one to access the other
    eg, listing fee
  3. Charge 3rd party advertising
  4. Charge producer/consumer for premium tools and services
    eg, for concierge-style service
  5. Charge consumer for access to high quality, curated producers
  6. Charge producer for change to signal high quality

NB: your platform should capture some value for every interaction (see the 5 above)

The 4 Cs

The 4C set of actions that producers and consumers need to do repeatedly for them to derive value from the platform are:  

On the producer side: 

1. Create — producers do this
Make it easy with neat content creation interfaces

2. Curate — the platform operator designs this
3 types:

  1. editorial
    1. useful for spotting patterns that can be automated and scaled
  2. algorithm
    1. detect good vs bad activity, based on rules
    2. watch out! Could lead to false positives and reject good creations
    3. scale carefully!
  3. social
    1. leverages networked & distributed nature of platforms
    2. give community tools: 
      1. voting
      2. flagging
      3. upvote/downvote
      4. … and aggregate this to sort and rank, and increase discoverability of the best producers
        1. useful as social proof
        2. useful for producers as feedback

On the consumer side:

3. Customize — the platform operator designs this (this is filter)

4. Consume — the consumer does this.
Encourage this with: newsfeeds, widgets

 

The platform stack: platforms are like sandwiches

The platform stack is: 

  • Tomato (top): network — marketplace — community
  • Avocado (middle): Enabling infrastructure (tools, services, rules)
  • Bread (bottom): the data layer
    This is the (learning) filter that powers relevance, solves the problem of abundance

Core take-out 3: the importance of OPP

Platforms are OPP. They're Open, Participatory, they're built for Plug & Play: it's key that it's easy for suppliers to plug-in to them. And they need to keep Pulling in the suppliers.

So, how can you design "pull"? How can you get suppliers to willingly, regularly produce core value units? 

Supplier relationship management — you could call it community management — becomes key.

How to design for producers

Think valuecosts, incentives — the value, costs, and incentives that matter to producers.

What are the set of actions that make it easy and simple (and fun) for producers to create core value units?

Reduce the friction. Give them access to consumers (without commensurate investment), and tools to make creating easy peasy. If possible, "break the skill barrier" — make them look good without making them work hard for it. (This is why people pay money to sit in restaurants with candles. Candles make us all look good.)

NB: you may have to give tools away. The money isn't in the tools (like it used to be: this isn't sell one thing and you're done). It's in the interactions (recurring revenue enabled by the web).

Build Cumulative Value — value that scales as the producer uses the platform more often, which will increase the repeatability and desirability of interactions. You can build cumulative value for the producer through: 

  1. Enhancing their reputation:
    1. explicitly, thru ratings and reviews 
    2. implicity, thru algorithms to promote them
  2. Enhancing their influence: 
    1. give them higher creation (and moderation) rights

Reward quality

Mitigate risk

(Also works for consumer too — cf Nir Eyal's Hooked model; he calls this "investment".)

"Scale the country club" — don't be for everyone! Act like a nightclub bouncer, and encourage the right quality and perceived quality of producers. Any additional friction should signal quality. If you add friction, ask: does it improve quality? Does it signal quality? Does it increase the repeatability of desirable interactions? (Also, is the interaction high value, high risk? In which case, trust is more key than ever.)

5 ways to onboard producers

  1. Bait
  2. FFF: Friction-free feedback
  3. Kill the time it takes to reach critical mass
    Find a hotbed of existing activity, where there's already an overlap of supply and demand
  4. Incentivize them!
  5. Staging: capture one side at a time
    1. eg, Opentable got restaurants (producers) on board first by providing restaurant management software (bait!) before any customers signed up
    2. NB: restaurants were highly fragmented tech-laggard inefficient vertical
    3. By providing management software, could aggregate table inventory, turn it into real-time data on table availability across restaurants… 
    4. By doing that, Opentable created (aggregated, accumulated) core value units… ready to sell through the core interaction
    5. NB: you can create fake data to kick start a community: many dating sites have done this

How to design for consumers

Think valuecostsincentives — the value, costs, and incentives that matter to consumers.

What are the set of actions that make it easy and simple (and fun) for consumers to consume core value units?

Curate: to "separate the award winners from the bathroom singers", the signal from the noise. 

3 ways to do this: 

  1. Editorial curation
    Useful to discover patterns that can be automated and scaled
  2. Algorithmic curation
    to detect good vs bad behaviours
  3. Social curation
    The best way to scale curation… and the best way to get to trust.
    1. Up/down-voting
    2. Rating
    3. Flagging
    4. Also useful signal (it's a form of value) for producers: motivates them to create

Build cumulative value: get them to invest in small actions and over time use these to improve the filters, and increase the value they get from your platform

Core take-out 4: to scale quality, better data, better filters

In order to Customize, a platforms filters have to work ever better. 

Ideally, your platform should encourage users to make small actions that increasingly improve the filter, so it's easier to customize the offer. (This way, despite abundance, you can still deliver relevance.)

The filter should take the core value unit and match it to what a consumer wants based on:

  • active intent
    eg, specific search
  • passive context
    eg, newsfeed, suggestions
  • point-in-time information
    ie, what's going on right now
  • cumulative data
    ie, based on your previous actions, behaviours
  • aggregated social data
    eg, "people like you also liked…"

How to get producers and consumers to stick around

However you define your key performance indicator, you'll want producers and consumers to stick around. 

Make sure long-term investment gives producers…

  1. greater reputation
  2. more influence

Make consumers stick around:

  1. Improve the filter so the matching is better
  2. Reward them for more engagement
  3. Give them variable rewards (cf Hooked)
  4. Re-activate deactivated users with off-platform notifications
  5. Activate them in the first place
  6. Set future triggers (cf Hooked)
  7. When onboarding (and when they see the spreadable unit — see virality below) show benefits 

Another core take-out: small but thriving is best

"Platforms that enable thriving interactions within a small user base scale much faster."

Facebook at Harvard is a great example. The reason? It isn't the size: it's the thriving interactions, which are key for virality, cycle time… and proof that the platform works!

 

Beware broken pipe dreams!

You might still be designing for a pipe world. 

Don't! Instead read this article in Wired. (I'll add the core take-outs here soon.)

How to design virality

  • Virality is key because: the ability to scale is a function of the current user base, so it accelerates of its own accord
  • "Virality is a business design problem, not a marketing or engineering effort. It requires design before optimization."
  • "Virality is user-generated scale."
  • Virality is key for a startup because:
  • "A startup is a company designed to grow fast… the only essential is growth" — Paul Graham
  • But beware! "Relying on virality as the only source of user-acquisition is a flawed strategy."
  • NB: virality is NOT DRIVEN by INVITES or Word of Mouth! It is about greater exposure thru use.
  • Platforms designed for platform scale rely on users sharing core value units with their network

Turn your users into marketers

  • This is why Instagram beat Hipstamatic: 
  • "Converting a single-user activity into a social, multi-user activity was the key reason for Instagram's growth
    • Both killed the skill barrier
    • Hipstamatic did it better
    • But Instagram encouraged users to post on Facebook
    • Instagram turned users into marketers
  • ake every time someone uses your product an instance of marketing
  • Think: repeatable design pattern

Designing the carrot

It's vital for viral that you design the carrot.

  • It's vital for viral that you design the right hooks & incentives to tell everyone every time they use it
  • It's vital for deep engagement that you design incentives to get people to stay on & engaged
    • Youtube users share videos, Kickstarter users share projects, Airbnb users share listings

The 4 elements of viral

4 elements of how a virus / a viral marketing meme spreads: 

  1. Host = hosts
  2. Air = external network
  3. Droplets = core spreadable unit
  4. Recipients

Viral step 1: the Host

Q: Why will sender send units out of the platform?

Don't forget your core!

This incentive MUST be aligned with the core interaction.

It must NOT distract from the core interaction, but be an integral part of it.

It should ENHANCE the VALUE that a user gets from the core interaction.
— think of Instagram, Kickstarter, Youtube

Most viral platforms:

  1. offer very low friction to create core value units
    1. KILL the skill barrier!
    2. note: what people spread is the content, not the platform
  2. high % of producers
    1. Which is why Whatsapp and Instagram spread quickly
  3. reward users with accelerating social feedback
    1. If you share off-platform, you get more upvotes… more exposure… more consumption… creates a VIRTUOUS CIRCLE
    2. Q: what's your platform's virtuous circle of production/consumption?
    3. Q: the host and the recipient should gain… and this should be contingent on the outcome

Viral step 2: Air = External network

Q: where will the unit from your platform meet current non-users?

Q: what is the network on which the unit spreads? 

Many use Facebook. For Instagram, this was Facebook. For Airbnb, this was Craigslist.

4 qs to consider: 

  1. choice of network is key
    1. Any network where people are implicitly or explicitly connected and allows (deliberately or not) you to insert your spreadable unit
    2. Q: which external network facilitates similar connections/interactions? Has a relevant look & feel?
    3. NB: Airbnb reverse engineered integration with Craigslist — despite Craigslist not encouraging integration by offering no APIs (Q: does this mean it could be possible via Whatsapp?).
      1. These sort of interactions were already happening on Craigslist
      2. Airbnb did it better, eg its use of reputation, trust, and its transaction mechanism
      3. Q: where are interactions like the ones you want to facilitate already happening?
  2. does this add value to the user on the external network?
    1. eg, PayPal added huge value to eBay users
    2. eg, Flickr was great for bloggers
    3. eg, Instagram connected your phone and its camera, with your Facebook page
  3. is there an under-used/not yet used network? First to use the network (or to unlock an element of the network) has an unfair advantage
  4. how easily can you integrate your spreadable unit into the network?

Viral step 3: Droplets = Core units

Q: what is the minimum transferable unit on the platform that can move onto an external network?

Q: are your core units designed to spread/be spread on an external network? (Hint: they should be!)
Think of Youtube videos, or the game state in a turn-based game.

"The value unit is a representation of the platform that can spread on an external network and act as a demonstration of the platform"

"The spreadable unit remains the most important, yet least understood element of designing for viral adoption."

How to design your core units:

  1. Make sure each one triggers an interaction on an external network
    1. eg, Youtube made sure its videos could be embedded on MySpace
  2. Think of the "producer-as-sender"
    1. encourage producers to spread their creation at the point of production
    2. make this part of the user workflow
    3. Q: When someone adds a message/reads a message, can they vote? like? love? and share!?
  3. "The spread of the unit should help to complete an incomplete interaction"
    1. eg, unanswered question = a spreadable unit that demands an answer (which is a form of social feedback)
    2. eg, "what shall we get Rose on her birthday?"
    3. eg, Survey Monkey needs respondents
    4. an incomplete interaction opens a gap, and creates an active call-to-action for recipient

Viral step 4: the Recipient

Q: why will the recipient do what you want them to do?

Q: why will a non-user on an external network convert to user on the platform?

One answer: incentivise conversions thru supporting content that travels with the unit

  • Should be a pitch
  • Should be a demo
  • Should contain a call-to-action
    • PS I love you. Get your free email at Hotmail.
    • Or implicit call-to-action to complete an incomplete interaction

SSCC — another way to think about viral

Send: maximise the outflow of units from the platform
Make sharing actions part of the creation workflow

Spread: ensure units spread on external networks
Ensure they can spill over from one network cluster to another
Create cross-cluster inventives
And then you'll benefit from network effects — when more production leads to more consumption
(Not sure what network effects are? Think of the fax machine or the telephone. The more people who have them, the more useful they become.)

Click: test pitches and call-to-actions to maximise clicks (and conversions)

Convert: minimise cycle time… make the steps/effort from the first exposure to a unit on an external network to the start of a new cycle as tiny and friction-free as possible

To plan this out, use the viral canvas!

 

How to scale quality: Create a Culture of Curation + Creation

It's all well and good getting lots of providers on your platform, but how can you make sure those are high quality providers and high quality core value units?

Step 1: encourage repeat participation
(that's the quantity bit)

Step 2: Curate

How?

  1. Editorial
    1. Editors will take on more abstracted roles
      eg, educate community on how to curate
      eg, Medium moved editorial function from centralized, non-scalable model to decentralised scalable model
  2. Algorithmic
    1. Editors identify repeat actions that can be automated
  3. Social
    1. Reputation model
      Give expert's opinions more weight than others
    2. Vote, rate, report
      Give badges, score, level, rating

Step 3: minimise/mitigate risk
eg, Airbnb and Uber invest in insurance & trust mechanisms

Step 4: Customise
Don't forget serendipity
Design a robust data acquisition strategy that captures consumer data during: 

onboarding 

ongoing, thru the lifecycle of the user