How do you grow an early-stage startup?

What is traction for a startup?

Traction is proof, in numbers, that people want what you are building. The venture firm CRV puts it simply: traction is what a founder can show instead of a forecast.

Traction is not likes, page views or how many people said "cool idea" at a meetup. Those can be early signs, but they do not prove anyone will use or pay for the product. If a number does not tie back to people using it or paying for it, it is not traction yet.

Free signups can fool you too. Theanna once had a free version. Nomiki Petrolla, Theanna's founder, shut it down within 45 days because it watered down the value, then doubled the price from $49 a month.

Before you launch, the venture firm Antler points to sign-ups, the share of visitors who sign up and a waitlist as early proof of demand.

• Revenue: money from customers, the strongest kind of traction.
• Active users: people who use the product again and again, the best number if you are not charging yet.
• Retention: how many customers stay month after month instead of leaving.
• Pre-revenue demand: a waitlist, pre-orders, or companies that agree to a paid pilot or beta.
• Growth rate: how fast any of these numbers is going up, week over week.

How much traction do you need to raise pre-seed or seed?

At pre-seed, you need less than most founders think. Hustle Fund, an early-stage fund, says that any revenue, users or email list above zero is exciting at this stage, and that it is fine to have no customers yet if you can show you are moving the business forward.

What pre-seed investors want to see is effort and learning. Hustle Fund looks for founders who have talked to dozens or even hundreds of possible customers and are running quick tests, like trying different landing page messages.

Seed is where hard numbers start to matter. Treat the benchmarks below as rough guides, not rules. A smaller number growing fast can beat a bigger number that is flat.

• Pre-seed

  What investors look for:

  Any revenue, users or email list above zero, plus proof you are talking to customers and running tests

  Source:

  Hustle Fund

• During an accelerator

  What investors look for:

  5 to 7% growth a week is good; 10% a week is exceptional; 1% a week suggests you have not figured it out yet

  Source:

  Paul Graham, Y Combinator

• Seed (B2B software)

  What investors look for:

  $500,000 to $1.5 million in annual recurring revenue for a strong round

  Source:

  CRV

• Seed (growth rate)

  What investors look for:

  Steady double-digit monthly growth is solid; low single-digit monthly growth raises questions

  Source:

  CRV

• Retention (B2B software)

  What investors look for:

  Monthly customer churn in the low single digits; net revenue retention above 100%

  Source:

  CRV

What are the 19 traction channels?

The 19 traction channels are the ways a startup can get customers, as listed in the book Traction by Gabriel Weinberg and Justin Mares. Brian Balfour's write-up of the book's Bullseye framework uses the same 19. Most founders only think about the channels they already know, and miss the one that would work best.

Many founders spend almost all their time building and hope people show up. Getting customers needs real time on your calendar every week, not whatever is left over. Peter Thiel, quoted in Brian Balfour's write-up of the framework, said poor distribution, not product, is the number one cause of failure.

• Viral marketing: users bring in other users.
• Public relations: coverage in news outlets.
• Unconventional PR: stunts and creative moves that get attention.
• Search engine marketing: paid ads on Google and other search engines.
• Social and display ads: paid ads on social sites and other websites.
• Offline ads: print, radio, TV, billboards.
• Search engine optimization: showing up in search results for free.
• Content marketing: articles, videos and guides people want to read.
• Email marketing: emails to people who opted in.
• Engineering as marketing: free tools that bring in customers.
• Business development: partnerships with other companies.
• Sales: reaching out to buyers directly.
• Affiliate programs: paying others to send you customers.
• Existing platforms: app stores, marketplaces and big sites where your buyers already are.
• Trade shows: booths and meetings at industry events.
• Offline events: meetups, dinners and workshops you run.
• Speaking engagements: talks at events your buyers attend.
• Community building: a group of users who help each other and spread the word.
• Targeting blogs: getting written up by the blogs and newsletters your buyers read.

How do I find which marketing channel works for my startup?

To find your channel, use the Bullseye framework: brainstorm an idea for every channel, rank them, pick your top three, run small cheap tests, then focus on the one that works. Here is how to do it step by step.

Theanna picked social media first, then partnerships, then in-person events. It has not done outbound sales yet, because there is not time to do every channel well.

1. 01

   Pick the one number you are trying to grow

   Choose one number that shows people want your product. Paul Graham says revenue is the best number to track, and if you are not charging yet, active users is the next best. Add one number for the value your product gives. At Theanna, it is the days from a founder connecting Stripe to their first paying customer: about 26.7 today, with a goal of 20. Write down where each number is today.

2. 02

   Brainstorm one idea for every channel

   Go through all 19 channels and write at least one real idea for each, even the ones that feel wrong. For each, guess what you would pay per customer, how many customers it could bring and how long a test would take.

3. 03

   Rank the channels into three groups

   Sort them into A (most promising), B (could work) and C (long shots). Look at what has worked for other companies selling to the same buyers you sell to.

4. 04

   Pick your top three

   Balfour's write-up says to choose an inner circle of about three channels, and to draw the line where your excitement clearly drops off. Testing a few at once is faster than one at a time, but testing too many means none get a fair shot.

5. 05

   Run a small, cheap test in each one

   Keep each test small. Balfour's example is running four ads, not forty, with a few hundred dollars. You only want a rough answer to three questions: what does a customer from this channel take to get, how many customers are there, and are they the customers you want?

6. 06

   Check your number every week

   Look at the same number on the same day each week and compare it to where it started. Write down what you tried and what happened, even when nothing moved.

7. 07

   Focus on the channel that works

   When one channel clearly beats the others, put most of your time into it and keep improving it. Keep going until it stops growing or gets too expensive, then run the whole process again to find the next one.

What should I measure for each traction channel?

Measure the one action that leads to a customer, not how many people saw your work.

Watch the right step, too. In a Meta ads test for a Theanna webinar, an easier form pushed signups from 148 to 653, but fewer than 1 in 10 people showed up. Then 24 people bought, and first-day sales of $6,578 beat the $5,994 ad spend. Nomiki says she was watching show rate when she should have watched ad spend per sale and what each customer is worth over time.

Search can work fast. Three weeks after Theanna's blog launched, Google showed it 53,200 times and sent 439 clicks, at an average spot of 6.9, which is page 1. But only 0.8% of people who saw it clicked, so the next test is better titles.

Here are common channels for early-stage startups and what to track in a test. Always tie it back to your main number, like revenue or active users.

• Sales (direct outreach)

  What to measure:

  Replies, calls booked, deals closed

  Good first test:

  Send personal messages to a short list of buyers who fit your ideal customer

• Content marketing

  What to measure:

  Visitors who sign up or book a call

  Good first test:

  Publish a few posts that answer questions your buyers already ask

• SEO

  What to measure:

  Search clicks to your site and sign-ups from search

  Good first test:

  Write one page for a search phrase your buyers use, then watch Google Search Console

• Email marketing

  What to measure:

  Open rate, click rate, replies, sign-ups

  Good first test:

  Send one useful email to your list with one clear next step

• Search ads

  What to measure:

  Price per click, price per sign-up, sign-ups

  Good first test:

  Run a small budget on a few search phrases with clear buying intent

• Social and display ads

  What to measure:

  Price per sign-up, sign-ups, paying customers

  Good first test:

  Run a few ads, not dozens, to a single landing page

• Community building

  What to measure:

  Active members, members who become users

  Good first test:

  Start or join one group where your buyers already talk

• Offline events and speaking

  What to measure:

  Conversations, follow-up calls, customers

  Good first test:

  Host one small event or give one talk, then follow up within a week

• Business development

  What to measure:

  Customers sent to you by each partner

  Good first test:

  Ask one company that serves your buyers to share your product

• Existing platforms

  What to measure:

  Installs or listings viewed, then sign-ups

  Good first test:

  List your product on one marketplace or app store your buyers use

Should I hire a growth person before I find a channel?

No. Test channels yourself first, so you know what works before you pay someone to do more of it. If you hire before that, you are asking a new person to guess, and you will not know if a bad result is the channel or the hire.

Stage 2 Capital, a fund that backs early sales teams, warns founders against stepping away from selling too soon. Early on, you need to learn firsthand what makes customers say yes. That knowledge is what you hand to your first hire.

When you do hire, match the hire to how you sell. Stage 2 says a self-serve company may need a growth hire first, and a company selling big deals to businesses may need a hands-on seller. Either way, show them the channel and the numbers that already work.

• Before hiring: you have tested a few channels and one clearly brings in customers.
• Before hiring: you track your main number every week and know where it started.
• Red flag: hoping a new hire will figure out who your customer is.

What does a weekly growth routine look like for an early-stage startup?

A weekly growth routine is: pick one number, run a few tests, check the number on the same day every week, and double down on what moves it. The founders who grow are usually the ones who keep it going week after week.

Most tests will not work. That is normal. Each one that fails tells you something, and you use it to plan a better next test. Paul Graham says you can use a growth target to make your decisions for you: if something helps you hit it, it is the right thing to work on.

Write each change down as a guess you can check. Nomiki guessed Theanna was losing customers because they did not know what to do next. She tested live onboarding sessions, and the share of customers leaving each month fell from 12% to 7%.

• Monday: look at your number and compare it to last week and to where you started.
• Monday: decide which tests to keep, stop or change based on what moved.
• Tuesday to Thursday: run the tests. Send the emails, post the content, make the calls.
• Friday: write down what you did and what happened, in one or two lines per test.
• Every few weeks: when a channel clearly works, move more of your time into it.