The First 100 Customers Course #17: How Matik Turned LinkedIn Alumni Into Its First 20 Customers

TL;DR

Matik’s first paid deal came from a dinner-party introduction. Founder Nik Mijic says the buyer tested the prototype, asked where to sign, and accepted a $2,500 price he chose on the spot. The next customers were less accidental: Mijic deliberately contacted former LinkedIn employees who had moved to companies without the internal presentation tools they already knew. He attributes Matik’s first 15 to 20 customers to that relationship-led motion.

The reusable mechanism is known workflow → person who remembers its value → new company missing it → warm conversation → paid proof → internal expansion. This is not a claim that Matik reached exactly 20 customers through one clean campaign. The milestone is founder-reported, the funnel is private, and funding helped answer enterprise trust objections. Copy the targeting logic, not the headline.

What you will build

You will build a portable-memory pipeline: a list of people who have already experienced the workflow your product replaces and now work somewhere that lacks it. The final output is not a giant contact export. It is a 30-person evidence list, a narrow message, a paid-proof offer, and a decision ledger that tells you when the motion is repeatable enough to extend beyond your network.

By day seven, you should have ten qualified names, five direct conversations, two concrete workflow reconstructions, and one paid next step or a clear stop decision. A warm reply is not validation. A familiar company name is not qualification. The pass condition is that a person can describe the old job, show its absence in the new account, and involve the owner of the problem.

Who this is for, and who should not copy it

This motion fits B2B software born from an internal tool, consulting workflow, agency process, implementation pattern, or repeated operator task. It is especially useful when the buyer needs context before a product demo makes sense and when the founder has fewer than 20 customers.

Do not use it as permission to scrape every former colleague. Do not imply endorsement from a previous employer, reuse confidential data, or ask a contact to buy simply because you once worked together. The method fails when memory is mistaken for demand. It also fits poorly when the product solves a new behavior nobody has experienced, the buyer is a consumer, or the purchase requires evidence your startup cannot yet provide safely.

Case snapshot and evidence limits

Stage	What the evidence supports	Limit
Problem origin	Mijic had worked on an internal LinkedIn system for data-driven presentations	No public measure of internal adoption or savings
Validation	Five to six months of customer, technical, and co-founder research	No interview count or conversion rate
First buyer	Dinner-party introduction, prototype test, $2,500 purchase	Buyer identity, contract length, and delivery effort are private
First 15–20	Personal network and former LinkedIn employees at new companies	Founder attribution, not an independently audited CRM cohort
Trust constraints	Survival and data security slowed early deals	No objection or loss percentages
Later expansion	Customer-success use spread to broader teams and sales	No net-revenue-retention figure

The complete SaaS Podcast transcript is the primary evidence for the first-customer sequence. Independent TechCrunch reporting confirms Matik’s 2019 inception, $3 million seed, $20 million Series A, product scope, and later named customers. It does not independently verify the first-20 attribution. Matik’s current site confirms that the company still automates presentations, documents, and reports, but current product breadth should not be projected backward onto its first sale.

The model: target portable memory, not generic similarity

Matik did not begin with “customer-success leaders at B2B companies” as a broad list. Mijic had a narrower signal. Some people had used an internal system at LinkedIn, left the company, and then felt the old manual work return. They did not need a category lecture. They already knew the job and could compare the new product with a remembered outcome.

1. Memory: the person has firsthand experience of the old solution or workflow.
2. Loss: the new company lacks that capability or performs it manually.
3. Authority: the person owns the job or can introduce its owner.
4. Proof: the founder can reproduce one useful output without a long deployment.
5. Trust: security, continuity, and implementation concerns are answered honestly.
6. Expansion: successful use reveals adjacent teams with the same underlying job.

Each stage filters the next. Prior experience removes explanation cost. Loss creates urgency. Authority converts curiosity into a process. Proof turns memory into a present result. Trust allows an early startup into a sensitive workflow. Expansion compounds inside an account. If one link is absent, the founder should stop and revise instead of adding more names.

Step 1: reconstruct the workflow before searching for people

Mijic’s wedge was concrete: sales and customer-success teams pulled data from systems such as Salesforce and Tableau, inserted it into presentation templates, and repeated the work for QBRs, renewals, and business cases. He did not begin with the feature “automated slides.” He began with a recurring job he had already observed.

Create a one-page workflow record with six fields: operator, trigger, inputs, manual steps, delivered artifact, and cost of delay. Add one example from the last 30 days. If the workflow happened only once, or if you cannot identify who owns its result, it is not ready for this motion.

Output: one workflow card in buyer language. Pass condition: three people independently recognize the same trigger and artifact. Stop condition: they understand the feature but cannot remember performing, buying, or suffering the job.

Step 2: build a portable-memory list

Start with the places where the workflow already existed: a former employer, client, community, agency, open-source project, or partner ecosystem. Then identify people who moved from that context into a company where the workflow plausibly reappears. The useful signal is the transition, not the old logo alone.

Field	Question	Keep rule
Prior context	Did this person use or own the workflow?	Direct evidence only
Transition	Did they move to a relevant new company?	Role or company changed
Current trigger	Is the same job likely present now?	One sourced signal
Authority	Can they buy or introduce the owner?	Named route to owner
Conflict	Would contact misuse confidential context?	Exclude any risk

Build 30 rows, then keep at most ten. This is deliberately small. A list that cannot survive manual verification is not a warm list; it is a loose audience with familiar employers.

Step 3: ask for reconstruction, not a demo

A portable-memory message should test whether the old pain exists in the new environment. It should not assume the person wants the product, and it should not borrow the former employer’s authority.

You worked with [workflow] at [prior context]. I am rebuilding the part that produced [specific output] without the manual [step]. Has that job followed you to [new company]? If yes, I would value 20 minutes to reconstruct the current version. No deck needed.

Metric: count replies that confirm the current workflow, not total replies. Pass condition: three of ten qualified people confirm the job and two involve the owner. Stop condition: replies are friendly but describe no current trigger.

Step 4: sell one paid proof

Matik’s first buyer tested the prototype and then asked to buy. Mijic had no order form and no established pricing model. He proposed $2,500, the buyer accepted, and the customer later expanded. The lesson is not to improvise every contract. It is to make willingness to pay observable before perfecting pricing.

Scope one output that can be delivered safely: one report, one dataset, one workflow, one team, or one recurring artifact. State what works, what is manual, which data is accessed, who approves release, and when the buyer can stop. Price high enough to cover founder time and implementation uncertainty.

Pass condition: the buyer pays, supplies real inputs, and uses the result in the target workflow. Stop condition: purchase depends on an unsupported security claim, a hidden service layer, or a roadmap promise you cannot reverse.

Step 5: treat trust as product work

Mijic says the first-ten objection was continuity: would Matik still exist in six months? Connecting to CRM and database information created a second constraint, security. Funding gave institutional credibility, while early security investment reduced data objections. Neither is a universal requirement, but both show that trust work belongs inside the acquisition system.

Build a trust packet before asking for a second enterprise deal: data-flow diagram, access boundaries, deletion policy, named incident contact, implementation plan, customer references you have permission to share, and a truthful continuity answer. Do not claim a certification you have not earned. If the buyer requires one, record it as a gate and decide whether the market justifies the cost.

Step 6: hand off only the pattern

After the first relationship-led customers, Matik added account executives, SDRs, inbound, and outbound. Mijic says the playbook remained iterative and the founder stayed in deals. He also recommends hiring comparable sales roles in pairs so a company can distinguish a person problem from a process problem. That advice reflects Matik’s funded context; a bootstrapped founder may need a cheaper test.

Before hiring, document six things: buyer, transition signal, trigger, proof artifact, top three objections, and next-step rule. Ask another operator to run ten accounts. If they cannot create qualified conversations without the founder silently repairing the context, the playbook is not ready. Keep the founder in weekly call review even after the handoff.

What failed, and what was merely premature

• No sales experience: the founders used advisors and repeated message tests; they did not solve it with a generic script.
• Continuity objection: early buyers saw startup survival as implementation risk. Funding helped Matik, but transparent scope and reversible deployment are the copyable parts.
• Security objection: sensitive integrations made trust structural, not a copywriting issue.
• Early sales hire: one early hire did not stay long. Hiring before the environment and role are clear can produce a false negative.
• Broad outbound: it came after the relationship-led wedge. The transcript does not show generic cold outbound creating the first 20.

Your seven-day implementation plan

1. Day 1: write the workflow card and one real before/after example.
2. Day 2: map 30 people with firsthand prior context and a relevant transition.
3. Day 3: verify current roles, triggers, conflicts, and route to the owner; keep ten.
4. Day 4: send ten reconstruction messages and prepare a question guide.
5. Day 5: run conversations; log current steps, frequency, data, owner, and cost.
6. Day 6: write one paid-proof scope with manual work and trust boundaries disclosed.
7. Day 7: ask for the paid next step and make a pass, revise, or stop decision.

Continue only if the same current job appears in at least three accounts and one buyer accepts a paid proof. Revise when the job is real but the owner, artifact, or trust boundary is wrong. Stop when prior familiarity produces polite calls without a present budget or consequence.

Lead Scorer implementation

Lead Scorer can support this motion when the source list is lawful, narrow, and already connected to a real workflow. It should not manufacture warmth. The operator still decides whether the prior relationship and current trigger justify contact.

1. Define context: use the icp-offer-context skill to store the workflow, qualified transition, disqualifiers, proof you may claim, and trust constraints. Output: one reusable ICP record. Human approval: confirm every claimed proof.
2. Separate audiences: create distinct lists for direct former colleagues, alumni you did not know, former clients, and adjacent lookalikes. Never mix relationship strength. Output: four small, tagged lists with provenance.
3. Score before enrichment: use a rubric weighted toward firsthand workflow memory, relevant move, current role, and sourced trigger. Keep only leads scoring at least 8/10. Stop when the prior context cannot be verified.
4. Spend contact credits last: run contact-discovery only for approved keepers who are about to be contacted. A known email is not permission to send.
5. Draft, then audit: create a draft campaign with one reconstruction message per person, then run outreach-qa-audit. Require a source for the old context and current trigger. Human approval remains mandatory before activation.
6. Close the loop: classify replies by confirmed workflow, timing, wrong owner, trust objection, or no current pain. Feed repeated objections into Content Studio as useful evidence, not as permission for an automated follow-up blast.

Build a ten-person portable-memory list for [workflow]. Keep only people with verified firsthand context, a relevant move, a current trigger, and a route to the budget owner. Score before enrichment. Draft one reconstruction message per keeper. Do not activate or spend contact credits without my approval.

Pass condition: every draft contains two verified signals and asks about the current job without presuming a relationship. Stop condition: the system relies on employer similarity, invented familiarity, confidential details, or unreviewed sending.

Saveable checklist

• One recurring workflow, one owner, one artifact
• Thirty candidates reduced to ten verified transitions
• Prior firsthand memory separated from generic alumni status
• Current trigger sourced before contact discovery
• Conversation asks for reconstruction, not a demo
• Paid proof states manual work, data access, approval, and stop rights
• Trust packet answers continuity and security honestly
• Founder-led pattern documented before a sales handoff
• Replies and objections update the ICP and future content
• No claim that relationship-led sales guarantees 20 or 100 customers

Sources and limits

• The SaaS Podcast, episode 336 with Nik Mijic: complete founder interview and the source for the $2,500 first sale, first 15–20 relationship-led customers, objections, and sales handoff.
• TechCrunch, October 13, 2021: independent confirmation of founding year, funding, product category, and named later customers.
• Matik official site: current product and security context.
• Andreessen Horowitz’s Matik investment note: interested-source context on the workflow and cross-team expansion.

No public source provides Matik’s early list size, response rate, demo conversion, CAC, ARR at customer 20, contract duration, or retention cohort. The thresholds in this course are operating rules for your experiment, not reported Matik benchmarks. “First 20” in the title rounds the founder’s stated 15–20 range and must not be read as an audited exact count.