dave mcclure startup metrics for pirates

5 minute version:

the point is NOT metrics. don't get too absorbed in metrics! the point is: build a feedback loop where you get data from users, analyze it (rigorously) and use that to change your product and your strategy.

dave mcclean: when you're gonna ship a feature: stop! first, test and see if the feature improves conversion! generally, additional features --> additional clutter --> lower conversion rate!. if you ship it, don't just move on to the next feature, measure the impact of the previous feature first!

you better love what you do, because you don't make much money. my first startup: i worked for 6-7 years; the startup sold for $500000 and i got $250000 (i.e. much lower salary than i would have made elsewhere).

i think startups should focus on three areas:

(these are more important than how you make money)

less metrics are better than you

metrics are very situation-dependent; there are no "best metrics" for everyone.

ppl tend to focus on # of unique visitors (top of the funnel) and # of paying subscribers (bottom of the funnel). but often the stuff in the middle (the conversion rates) are more important.

the lean startup church of metrics:

if any of you want a meeting with me, don't talk about that, but what i want you to show me is: 3 conversion metrics based on your user funnel, how they've changed over the last month or two, what experiments you've run and whether they've succeeded or failed, and how this has affected your product or marketing.

user experience and distribution is more important than engineering

need to minimize time thru the code/measure/learn loop!

practices to help u code faster:

            Unit Tests
            Usability Tests
   Continuous Integration
 Incremental Deployment
     Free & Open-Source
        Cloud Computing
  Cluster Immune System
   Just-in-time Scalability
      Developer Sandbox
 Minimum Viable Product

practices to help you measure:

Funnel Analysis SplitTests? Cohort Analysis ClearProductOwner? Net Promoter Score Con?nuousDeployment Search Engine UsabilityTests? Marketing Real?meMonitoring Real-Time Alerting CustomerLiaison? Predictive Monitoring

practices to help you learn:

LearnFaster? Split Tests Customer Interviews Customer Development Five Whys Root Cause Analysis Customer Advisory Board Falsifiable Hypotheses Product Owner Accountability Customer Archetypes Cross-functional Teams Semi-autonomous Teams Smoke Tests

sean ellis: don't spend money on marketing, don't try to grow , until you have, and understand, a core base of dedicated users. don't go viral until you don't suck. you might spend a lot of time with a small base of customers (~100 customers, and/or a small marketing campaign with 100 clicks a week). "start from a base of people who actually use your crappy product", ideally a niche market without much competition so that you can start with a really crude product. before growing, you don't want to be a 6/8, you want to be an 8/10.

((as David Skok would say, don't hit the "accelerator pedal" until you have a viable business model))

when looking at marketing channels, don't spend a lot of time optimizing a marketing channel if its total size ((est. size before saturation)) is small. i.e. if one channel might bring you a thousand users, and the other one a million, then spend a lot of time optimizing the big channel until it is cost effective.

simple is better: fewer metrics are better than many (("KPI"))

note: once you have conversion rates, you can calculate a value for each user at each discrete stage in the funnel

ship early, ship buggy; want to find a niche with little enough competition so that you can be a 6/10 on first ship ((rework disagrees with shipping buggy))

you gotta experiment with website copy, images, and keywords. you can design experiments that don't convert and then you don't have to pay for it

website design tips:

if ppl hate what you've done, that's better than if they feel neutral. b/c if they hate what you've done, chances are it's similar to something they love.

you want to find the 1-3 features that matter. one way to look: every week, take a feature away, and see if users scream (if they do, bring it back, only better)

his biz model. priorities:

  1. 1 drive usage (activation, retention; note: retention is usually more important than referral, although sometimes it's possible to lots of users who don't engage and drive usage, but that refer you to users who do, but that's)
  2. 2 acquire users (this comes first chronlogically but is less important than #1)
  3. 3 make money (obviously you need to do this eventually but worry about it after the other two)

in other words:

  1. 1 find a problem
  2. 2 figure out the conversion
  3. 3 see if they come back

THEN start to think about distribution, and cost structure, THEN think about optimizing for revenue

in other words:

  1. 1 find a landing page that converts
  2. 2 see if they come back

if ur into bootstrapping, may want to figure out conversion first, then make money, and only then to drive usage

in general, should not be trying to do 2 of these at the same time

3 core models:

example 1-page biz model: for a video sharing site, types of customers might be "visitor(watcher), contributor(uploader), distributer(linker)". stages: activation, retention, revenue. types are cols, stages are rows. in each box, an operationalized definition of success for that stage for that customer type. e.g.

Watch a slideshow Submit a slideshow Share/Embed a slideshow
Visit 1x/mo. for 3+ mo’s Submit “hi-qual” ppt (=> 1K views/yr) every mo.Share/Embed => drive 100 visits/mo for 3+ mo’s
Revenue Ad Click; Qualified Lead Generated $/mo Premium Sub;Pay $$+ for traffic or leads Drive 10+ Premium users, 1K+ Viewers

ur product must be:

optimization problem. want to optimize two different things:

note: a metric is really a strategy with local maxima discovery

not a lot of time of feature dev, but in fact you should be spending more time on optimizing existing features

estimate cost, and, separately, value, of a customer at each stage in the funnel

don't JUST look at LTV, look also at value for the first transaction or the first month of transactions, so that you aren't overly optimistic

" Marketing Plan = Target Customer Acquisition Channels 3 Important Factors = Volume (#), Cost ($), Conversion (%) Measure conversion to target customer actions Test audience segments, campaign themes, Call-To-Action (CTAs)

[Gradually] Match Channel Costs => Revenue Potential Increase Vol. & Conversion, Decrease Cost, Optimize for Revenue Potential Avg Txn Value (ATV), Ann Rev Per User (ARPU), Cust Lifetime Value (CLV) Design channels that (eventually) cost <20-50% of target ATV, ARPU, CLV

Consider Costs, Scarce Resource Tradeoffs Actual $ expenses Marketing time & resources Product/Engineering time & resources Cashflow timing of expense vs. revenue, profit "

delegate each metric to someone to own

"one step at a time (product vs market vs revenue):

Make a Good Product: Activation & Retention Market the Product: Acquisition & Referral Make Money: Revenue & Profitability " (if ur bootstrapping, making money comes before marketing)

good engineering is valuable b/c it shortens cycle time

3 stages of company/3 times for investments:

" Investment #1: Incubate (“Product”)

Structure 1-3 founders $25K-$100K investment Incubator environment: multiple peers, mentors/advisors

Build Functional Prototype / “Minimum Viable Product” (MVP): Concept->Alpha, ~3-6 months Develop Minimal Critical Feature Set => Get to “It Works” Instrument Basic Dashboard, Conversion Metrics Test Cust. Adoption (10-1000 users) / Cust. Satisfaction (Scale: 1-10)

Demonstrate Concept, Reduce Product Risk, Test Functional Use Develop Metrics & Filter for Follow-on Investment

Investment #2: Seed (“Market”)

Structure 2-5 person team $100K-$1M investment Syndicate of Angel Investors / Small VC Funds

Improve Product, Expand Market, Test Revenue: Alpha->Beta, ~6-12 months Customer Sat ≥ 6 => Get to “Doesn’t Suck” Setup A/B Testing Framework, Optimize Conversion Test Marketing Campaigns, Cust Acqstn Channels

Prove Solution/Benefit, Assess Market Size Test Channel Cost, Revenue Opportunity Determine Org Structure, Key Hires

Investment #3: Venture (“Revenue”)

Structure 5-10 person team $1M-$5M investment VC Investors

Make Money, Get to Sustainability: Beta->Production, 12-18 months Customer Sat ≥ 8 => “It Rocks, I’ll Tell My Friends” MktgPlan? => Predictable Channels / Campaigns + Budget Scalability & Infrastructure, Customer Service & Operations Connect with Distribution Partners

Prove/Expand Market, Operationalize Business Future Milestones: Profitable/Sustainable, Exit Options


btw, don't pitch me (or email me): if you want to talk to me,

  1. 1 read my stuff
  2. 2 get a referral from an investor or a portfolio company
  3. 3 be concise, don't suck

Additional References: Influence: The Psychology of Persuasion Robert Cialdini (book) The Mating Mind Geoffrey Miller (book) Putting the Fun in Functional Amy Jo Kim (etech 2006 preso) Futuristic Play Andrew Chen (blog) Don’t Make Me Think Steve Krug (book) Designing for the Social Web Joshua Porter (book, website) Startup Lessons Learned Eric Ries (blog) Customer Development Methodology Steve Blank (presentation, blog) Sean Ellis (blog) Hiten Shah / Neil Patel (website) How To Pitch a VC Dave McClure? (slides, NSFW) Understanding Comics Scott McCloud? (book)

example of what not to do with metrics: have too many and have a big spreadsheet that is hard to get a sense of, for example, one whose columns are dates, and whose rows are:



site usage:

Appendix (in slides only)


" What do users do on their first visit? "


Activation Criteria: 10-30+ seconds 2-3+ page views 3-5+ clicks 1 key feature usage

do LOTS of landing page & A/B tests – make lots of dumb guesses & iterate FAST "

" Example Activation Goals Click on something! Account sign up / Emails Referrals / Tell a friend Widgets / Embeds Low Bounce Rate

Activation Tips Less is more Focus on user experience / usability Provide incentives & call to actions Test and iterate continuously

Key Metrics to Track Pages per visit Time on site Conversions

Tools Crazy Egg (Visual Click Mapping) Google Website Optimizer (A/B & Multivariate Testing) (B2B Lead Generation Management)

Resources Experimentation and Testing: A Primer Landing Page Design Toolbox: 100 Tips & Tools Landing Page Tutorials & Case Studies 101 Easy Easy to use Google Website Optimizer



" How do users come back? How often? "

" Automated emails: lifecycle emails @ +3, +7, +30d status / “best of” weekly/monthly “something happened” emails BUT: make it easy to unsubscribe

Tip on emails: > 80% or more on SUBJECT LINE < 20% or less on BODY TEXT


" Cohort Analysis: Distrib of Visits over Time Rate of Decay Effective Customer Lifecycle

Retention Methods Automated Emails

" Example Retention Goals 1 - 3+ visits per month 20% open rate / 2% CTR High deliverability / Low spam rating Long customer life cycle / Low decay Identify fanatics and cheerleaders

Retention Tips Email is simple and it works BUT make unsubscribe easy 80% subject line / 20% body text ACTUALLY 99% subject line / 1% body text Fanatics = virality + affiliate channel (bloggers?)

Key Metrics to Track Source Quantity Conversions Visitor Loyalty Session Length

Tools Campaign Monitor / MailChimp? (email newsletter software) / TriggerMail? (site-centric email management) Litmus (email and website design testing - clients / browsers)

Resources 30 free HTML email templates Best Practices in Writing Email Subject Lines Learning Viral: Viral Emails of "

Acquisition " Where are users coming from? "

" Acquisition Methods SEO / SEM Blogs Email Social Media & Social Networks Domains

Top 10 - 100 words Your Brand / Products Customer Needs / Benefits Competitor’s Brand / Products Semantic Equivalents Misspellings

Things to analyze Sources Volume Cost Conversion

Key Metrics to Track Quantity (#) Cost ($) Conversions (%)

Tools Google Analytics (web analytics) Google Keyword Tool (keyword research tool) SEO Book Tools (SEO related tools)

Resources SEO Book Blog The Social Media Manual: Read Before You Play Strategies to ruthlessly acquire users "


"How do users refer others?"

" Focus on driving referrals

Referral Methods Send to Friend: Email / IM Social Media Widgets / Embeds Affiliates

Viral Growth Factor = X * Y * Z X = % of users who invite other people Y = average # of people that they invited Z = % of users who accepted an invitation

Tools Gigya (social media distribution & tracking tool) ShareThis? / AddThis? (sharing buttons) / GetMyContacts? (PHP contacts importing & invitation software)

Resources Seven Ways to GO VIRAL What’s your viral loop? Understanding the engine of adoption Metrics: Where Users Come From "


" Revenue Tips Don’t Rely on AdSense? (only) Start Free => 2% “Freemium” Subscription / Recurring transactions Qualify your customers -> Lead generation (arbitrage) Sell something! (physical or virtual)

Resources & Tools Revenue Metrics (Andrew Chen) How to Create a Profitable “Freemium” Startup (Andrew Chen) 2008 Affiliate Marketing Review (Scott Jangro)


note: the slides have pics of metrics dashboards in the appendix

some of the other slide presentations have a big list of types of marketing channels and their characteristics

(not that important:

measurement components:

when measuring referral, measure not just # of referrals, but # of referrals that convert

example metrics (/conversion rate/value (not cost)):


Visit Site (or landing page, or external widget) 100% .01

Doesn't Abandon (views 2+ pages, stays 10+ sec, 2+ clicks) 70% .05


Happy 1st Visit (views X pages, stays Y sec, Z clicks) 30% .25

Email/Blog/RSS/Widget Signup (anything that could lead to repeat visit) 5% $1

Acct Signup (includes profile data) 2% $3


Email Open / RSS view -> Clickthru 3% $2

Repeat Visitor (3+ visits in first 30 days) 2% $5


Refer 1+ users who visit site 2% $3

Refer 1+ users who visit site 1% $10


User generates minimum revenue 2% $5

User generates break-even revenue 1% $25