# SaaS startup marketing playbook

> A week by week plan for the first 100 customers: manual outreach, narrow ICP, one channel, and the activation work that makes the next 100 cheaper.

Source: https://saas-marketing.net/playbooks/saas-startup-marketing/
Topic: SaaS Marketing
Type: playbook
Published: 2026-09-11
Last updated: 2026-09-11
Publisher: SaaS Marketing (saas-marketing.net)
License: CC BY 4.0. Quote or republish with attribution and a link to https://saas-marketing.net/playbooks/saas-startup-marketing/

## Short answer

The first 100 SaaS customers come from manual work, not campaigns. Spend weeks 1 and 2 narrowing the ICP and running 20 customer interviews, weeks 3 and 4 rewriting the message and homepage around what you heard, weeks 5 to 8 running a single channel chosen by your ACV, and weeks 9 to 12 fixing activation and collecting proof. Pick the channel the founder can personally sustain, because consistency beats reach at this size.

## Key takeaways

- A 12 week sequence beats a channel list because the first four weeks decide whether the next eight can work
- Twenty customer interviews cost two weeks and prevent six months of writing to the wrong buyer
- Choose one channel by ACV: outbound above $15k, community and content below, paid almost never before 100 customers
- Seed stage CAC payback of 4 to 5 months is achievable because the acquisition work is founder time, not spend
- Activation work in weeks 9 to 12 is what makes customers 101 to 200 cheaper than the first hundred
- Switching channels at week 6 because nothing happened yet is the single most common way this plan fails

---

Almost every early stage marketing plan fails the same way. The founder picks four channels because a blog post said to diversify, gives each one a third of the effort it needs, and by month four has a thin blog, a dead LinkedIn account, a Google Ads bill and forty unanswered cold emails. Nothing failed, exactly. Nothing got enough attention to prove anything either. The plan below does one thing at a time in an order where each step makes the next one cheaper.

## Why the first 100 customers come from work that does not scale

Manual work is how you buy information you cannot get any other way. Sending 200 hand-written messages teaches you which objection repeats, which segment replies fastest and which word in your description makes people flinch. A campaign that scales cannot teach you that, because scaling requires you to already know it.

Stripe is the canonical example. In the early days the founders would offer to set the product up on a prospect's machine during the conversation rather than sending a link and hoping, a practice that got named the Collison installation. It did not scale. It told them exactly where integration broke, which is what made the docs and the API the product's best marketing asset for the next decade.

Superhuman is the other useful case. Rahul Vohra ran every early user through a personal onboarding call and used the Sean Ellis survey question, tracking the share of users who would be very disappointed to lose the product, with 40% as the threshold that signals fit. Manual onboarding was not a stopgap. It was the measurement instrument.

Unscalable work is not a phase you suffer through until real marketing starts. It is the research that determines which channel will scale later, and skipping it means guessing about the one decision that costs the most to get wrong.

The rest of this page is a 12 week sequence for a pre-Series-A team. It assumes one founder with roughly fifteen hours a week and a budget under $4,000 a month. For the version with a marketing hire already in place, [SaaS company marketing by stage](/guides/saas-company-marketing/) picks up where this stops.

## Weeks 1 and 2: narrow the ICP until it is uncomfortable

Cut the target market until naming a specific company takes you two seconds. "B2B software companies" is not an ICP. "Series A to B2B SaaS companies with 8 to 30 sales reps running territory rules in a spreadsheet" is one, because you can list forty of them by Friday.

Then run 20 interviews. Split them: eight with customers or users you already have, seven with people who fit the profile and use a competitor, five with people who evaluated something in this category and chose to do nothing. That third group is the most useful and the one everybody skips, because their objection is the one your funnel is losing to.

Ask about the last time the problem actually hurt. Dates, tools, who complained, what broke. You are listening for trigger events, because campaigns get built around triggers and not around personas.

**Weeks 1 and 2, in order**

Two weeks is the budget. Founders who stretch this to six are usually avoiding the harder step that comes next, which is committing to a segment and turning away revenue that does not fit it.

## Weeks 3 and 4: rewrite the message, then the homepage

The message comes from the verbatim document, not from a positioning workshop. Take the three phrases that repeated across five or more calls and build the homepage around them, in the customer's vocabulary rather than your category's.

Most early homepages describe the product. The ones that convert describe the situation the buyer is in immediately before they go looking. If five interviews described a spreadsheet breaking before a board meeting, the headline is about the spreadsheet, not about "unified revenue operations".

Rewrite in this order, because it maps to how a visitor reads: headline that names the buyer's situation, subhead that says what replaces, one proof element above the fold, three specific outcome statements, pricing visible or an honest reason it is not, one call to action repeated. Cut every sentence that would survive unchanged on a competitor's site.

| Element | Common early version | The version that converts | Why |
| --- | --- | --- | --- |
| Headline | "Revenue operations, simplified" | "Stop rebuilding territory rules in a spreadsheet every quarter" | Names the trigger, not the category |
| Subhead | "The all in one platform for modern teams" | "Routing rules that update when your headcount does, live in 20 minutes" | Says what changes and how fast |
| Proof | Six grey customer logos | "Used by 14 RevOps teams at Series A to C B2B SaaS companies" | Specific and checkable beats prestigious and vague |
| CTA | "Get started" | "See it on your own routing rules" | Matches the sales motion you actually run |

Ship it in week 4 even if it is imperfect. You will rewrite it again at 50 customers, and the version after that will be better because of what the outreach in weeks 5 to 8 teaches you.

## Weeks 5 to 8: one channel, chosen by ACV

One channel. The choice is set by your average contract value, because ACV determines how many hours of human effort each customer can justify.

The tiebreaker matters more than the table. A founder who hates writing will not publish weekly for eight weeks regardless of what the ACV says, and an abandoned channel teaches you nothing. Pick the one you will still be doing in week 7 on a bad week. The full ranking with costs and time to first result is in [SaaS marketing channels, ranked](/guides/saas-marketing-channels-ranked/).

Whatever you pick, define the weekly output as an action you control. "Book six demos" is an outcome and it will make you feel like a failure in week 2. "Send 25 researched messages" is an output and it is the thing that produces the outcome six weeks later. Track outputs weekly, outcomes monthly.

If outbound is the row you landed on, the mechanics of list building, sequencing and qualification are covered properly in [lead generation for a SaaS startup](/guides/saas-startup-lead-generation/). If you chose the search row, start small: [seed stage SaaS SEO](/playbooks/seed-stage-saas-seo/) is deliberately a narrow programme, because a full content operation at this stage is a way of looking busy.

Six weeks in, nothing has obviously worked, and the temptation to switch channels is overwhelming. Almost every channel has a lag longer than six weeks. Switching now means you never learn whether any of them work, and three quarters later you have four dead channels and no data. Give the channel its full eight weeks, then judge it on conversations created, not revenue.

## Weeks 9 to 12: activation and proof, so the next 100 cost less

Acquisition gets the attention and activation decides the economics. If 60% of signups never reach the moment the product becomes useful, every pound spent on the front end is wasted at a rate that no channel optimisation can fix.

Define your activation event as a specific action inside a specific window. Not "logged in twice". Something like "connected a data source and ran one report inside seven days", chosen because customers who did it renewed and customers who did not, churned. Then instrument it and watch the drop-off week by week.

The fastest wins in this four week block are usually unglamorous. Fix the three setup steps where people stall, write the four onboarding emails that walk someone to the activation event, and add an in-product checklist. Teams routinely find 10 to 20 points of activation improvement from that alone, which is worth more than a new channel.

Write the four emails around the blocker rather than the feature. Email one, sent within ten minutes of signup, does one thing: gets them to the single setup step that everything else depends on. Email two, at 48 hours, goes only to people who did not complete it and offers to do it for them on a call, which sounds absurd at scale and works perfectly at 30 signups a month. Email three shows one customer's actual configuration. Email four, at day six, asks a question that gets a reply, because a reply tells you why someone stalled and no analytics tool will.

Do the onboarding calls yourself for the first hundred customers. Linear ran an invite-only beta for months and controlled who got in; Slack's early team watched individual teams use the preview and fixed things by hand. Both look like customer support from the outside and both were product research. The value is not the call, it is that you hear the same confused sentence four times in a fortnight and finally understand what your homepage is failing to say.

Proof collection runs in parallel. Every customer who reaches activation gets asked for one of three things: a two sentence quote, a specific number, or thirty minutes for a case study. Ask within a week of their first success, because gratitude has a short half-life.

**40%** Share of users who would be very disappointed to lose the product, the survey threshold Superhuman used as a product market fit signal

Run that survey at the end of week 12 with everyone who activated. Under 25% means go back to weeks 1 and 2 rather than forward to scaling, and that answer is worth more than another quarter of outbound.

## The weekly scorecard and what good looks like at seed

Six numbers, one page, reviewed every Friday for fifteen minutes. Anything more elaborate at this stage becomes a reporting job that nobody has time for.

| Number | Week 4 target | Week 8 target | Week 12 target |
| --- | --- | --- | --- |
| Researched messages or assets shipped | 40 | 140 | 220 |
| Reply rate on outreach | 15% | 25% | 30% |
| Qualified conversations booked | 4 | 14 | 26 |
| Signups or trials from all sources | 6 | 20 | 38 |
| Activation rate within 7 days | Baseline measured | Baseline plus 5 points | Baseline plus 10 to 20 points |
| New paying customers | 1 to 2 | 5 to 9 | 12 to 20 |

Those targets assume a $8,000 to $25,000 ACV B2B product with one founder on the work. Halve them for a two-day-a-week founder and double the customer count for a self serve product at $50 a month. Broader stage-by-stage figures sit in the [SaaS marketing benchmarks](/research/saas-marketing-benchmarks/).

Seed stage economics are unusual and worth saying plainly. CAC payback of four to five months is achievable here, not because the marketing is brilliant but because the acquisition cost is mostly founder hours that the P&L already carries as salary. That number will get worse when you hire, and a board that anchors on it will be disappointed at Series A.

## What this costs, and the three ways it fails

Budget is small and time is not. Expect $1,500 to $4,000 a month in real spend and fifteen founder hours a week, which is the expensive half.

The honest cost line: CRM and email tooling $100 to $400 a month, a data or enrichment tool for outbound $100 to $300, design and production help $800 to $2,000, and a domain plus site build somewhere between $0 and $4,000 once. A [SaaS marketing budget template](/templates/saas-marketing-budget/) makes the tradeoffs concrete before you commit to a tool contract.

Three failure modes account for most of the plans that die.

- Widening the ICP after two quiet weeks, which resets the interview learning and makes every message generic again
- Running two or three channels at a third of the effort each, so none reaches the point where it produces signal
- Hiring a marketer at week 10 to rescue a plan that is only halfway through its first cycle

The third one is worth dwelling on. A hire at week 10 inherits a message that has not been validated and a channel with no data, and will reasonably start over. You lose the quarter. Wait until one channel has produced customers for three straight months before adding anyone, and when you do, the [first 90 days as a SaaS marketing lead](/checklists/first-90-days-saas-marketing/) is the handover document.

This gets the machine working. It does not get you to 100 customers. For a $5,000 to $20,000 ACV product, 100 customers usually takes nine to eighteen months, and the first 20 take nearly as long as the next 80 because you are still learning who to sell to. Plan the runway on that timeline, not on the 12 weeks.

## What to do in week one

Book five customer conversations before you do anything else, including reading further. Write your exclusion list on Monday and use it to disqualify one inbound lead by Friday, because that is the only proof the ICP is real.

Pick your channel in week 5 using the ACV table, then commit to the weekly output number in writing where a cofounder can see it. Do not evaluate the channel before week 8. When you review it, judge conversations created rather than revenue, because revenue lags by a full sales cycle and will tell you to quit something that is working.

At week 12, run the survey, read the activation numbers, and choose deliberately between another eight weeks in the same channel and adding a second one. Most teams should stay. The companies that compound tend to have run one channel for two years, which is the pattern the whole of [SaaS marketing](/saas-marketing/) keeps confirming, and the content side of that commitment is mapped in [content marketing for SaaS companies, by stage](/guides/content-marketing-for-saas-companies/). For the strategic layer underneath this playbook, [B2B SaaS startup marketing](/guides/b2b-saas-startup-marketing/) covers the positioning and pricing decisions this plan assumes you have already made.

## Frequently asked questions

### How do you get your first 100 SaaS customers?

Through manual, unscalable work. Narrow the ICP to a segment you can name, interview 20 of them, rewrite the message around their words, then run one channel for eight weeks with a weekly output you can sustain. Most seed stage companies reach 100 customers through direct outreach and community participation before any campaign scales.

### How much should an early stage SaaS startup spend on marketing?

Seed stage SaaS typically runs marketing at 20% to 30% of ARR, but at very low ARR that percentage is meaningless. A more useful floor is $1,500 to $4,000 a month covering tooling, contract design and production help, with founder hours as the largest real cost. Paid acquisition before 100 customers is usually premature.

### What marketing channel should a SaaS startup start with?

Pick by average contract value. Above roughly $15,000 ACV, direct outbound to named accounts works because each customer justifies hours of effort. Between $3,000 and $15,000, community participation and bottom of funnel content tend to win. Below $3,000, self serve requires volume, so search and integration listings matter earlier. Always pick the one the founder can sustain weekly.

### How long does it take to get 100 SaaS customers?

For a $5,000 to $20,000 ACV B2B product, nine to eighteen months is typical, with the first 20 taking nearly as long as the next 80. Self serve products at low price points can move faster in customer count but take longer to reach meaningful revenue. A 12 week plan gets the machine working, not to 100 customers.

### Is doing things that do not scale still good advice for SaaS?

Yes, and it is research rather than a phase to rush. Manual onboarding, hand written outreach and founder led demos tell you which objections repeat, which setup steps break and which segment converts. That information is what makes the scalable version work later. Stripe famously set up the product on customers' machines during early sales conversations.

### What metrics should a seed stage SaaS track weekly?

Six numbers: qualified conversations booked, signups by source, activation rate within seven days, new paying customers, reply rate on outreach, and one channel specific output number such as posts published or communities answered in. Track them weekly on one page. Revenue targets are too slow to steer by at this stage.

### Should a SaaS startup do SEO before product market fit?

Start a small amount early and expect nothing for two quarters. Publishing comparison pages, integration pages and one or two problem focused articles a month costs a few hours and compounds, but SEO cannot be your only channel before 100 customers because the feedback loop is too slow to teach you anything about your buyer.