1. The Play Library

The specific things you can run, from easiest to hardest.

A play is one specific reason to reach out to one specific group of people. "Email our closed-lost deals from last year about the feature we just shipped" is a play. "Do outbound" is not.

Good outbound is a series of plays, not one big campaign. You run one, read the results, and pick the next.

The plays below are grouped by how well the recipient knows you, which turns out to be the same thing as how hard the play is and how much setup it needs. Group 1 you can run this afternoon. Group 3 needs a few weeks of preparation. Work top to bottom. Every play in groups 1 and 2 sends from your normal work email and needs no special setup.

What Lightfield is doing behind all of these

Two things, and it is worth knowing which is which.

It finds and screens companies using the description of your ideal customer, built from your own won and lost deals. A person screening 300 companies gets sloppy by company 40. Lightfield applies the same standard to all 300, and it can check things a person would not bother to look up, like recent funding or who they are hiring.

It finds the right people inside those companies using the description of your buyer. Not the most senior person, and not whoever is easiest to find. The kind of person who actually bought from you before.

Both descriptions come from the Setup: Your Knowledge Base article. Build them first. Most of the plays below are just these two capabilities pointed at a different starting group.

Group 1: people who already know you

The fastest pipeline available to you, and almost always the most overlooked. These people recognize your name, so they reply at much higher rates than strangers. All of these send from your normal work email. No setup, no new domains, nothing to buy.

Revive

Go back to deals you lost, opportunities that stalled, customers who churned, and people who signed up and never came back.

When to run it: first, before anything else in this playbook.

The version that works best: you lost deals because you were missing something, and you have since built it. Pull the deals that died on that specific gap and go back with a real reason. Different companies wanted that missing feature for different reasons. One wanted it to save time, another to satisfy an auditor, another to stop losing customers. Group them by the reason, not by the feature, and write to each group about their reason.

Build a list of closed-lost deals from the last 18 months where the loss reason relates to [feature]. Group them by the specific problem they were trying to solve, write a separate sequence for each group, and include 2 to 3 contacts per company.

Why it works: you are not asking them to reconsider a decision. You are telling them the reason for their decision no longer applies.

Contact more than one person. Two or three people per company. The person you dealt with last time may have left, or changed jobs, or simply be on holiday. If you bet the whole thing on one inbox you lose the company to bad luck.

Expand

Reach new people at companies that are already your customers.

When to run it: second. It is the shortest path to revenue you have, because the hard part, convincing a company you are worth paying for, is already done.

The part most people miss: your call recordings and notes are full of problems your customers described out loud and you never followed up on. Lightfield can go find them.

Look across our customer accounts and find problems or needs that came up on calls in the last 12 months that match something we do but they are not using.

Champions who moved

Find people who liked working with you and have since moved to a different company. Lightfield can find where they went.

When to run it: set a regular rhythm, monthly or quarterly. People change jobs continuously, so this refills on its own.

Why it works: someone who used your product and had a good experience carries that with them into a company where nobody has heard of you. They already know it works, they already know how to explain it internally, and they usually have budget influence in their new role. A cold email to that company would be a two percent shot. An email to them is not.

Your own product users

If you have a free tier, a self-serve plan, or a second product, the people already using it are prospects who have used your software.

When to run it: as soon as your user data is in Lightfield. This is often the single largest warm audience a product company has, and it is almost always unworked.

How it goes: ingest your users, score their companies against your ideal customer description to find the ones big enough or right enough for your paid or enterprise tier, then sequence them. You already know which features they use and how often, so the message can be specific in a way no cold email ever could be.

We ingest all our product users into Lightfield. Score their accounts against our enterprise ICP, then build sequences to reach the top ones, using what we know about how they use the product.

Existing LinkedIn connections who went cold

People you are already connected to on LinkedIn and have not spoken to in months.

When to run it: any time. This one is easy to overlook because it does not feel like outbound.

Why it works, and why it is cheap: you need no connection requests at all, so it does not touch your weekly LinkedIn limit, which is the scarcest resource in outbound. A message to an existing connection is a conversation inside a relationship rather than a request from a stranger, and LinkedIn treats it far more permissively.

One mechanical thing to know: you cannot send a connection request to someone you are already connected to, so build this as a DM-only sequence with no invite step. Teams running this at real scale see it work well.

Ask for referrals

Go to happy customers and to your investors and ask for specific introductions.

Why the specificity matters: "do you know anyone who might need us" gets almost nothing, because you have handed the other person homework. "You mentioned you know the VP of Engineering at [Company], would an introduction make sense" gets a yes or a no. You did the work of identifying who; all they have to do is decide. Lightfield can find the specific people worth asking about by looking at who your customers and investors are connected to.

Group 2: people one step away

They half know you, or there is a genuine reason for your message to exist. Reply rates are lower than group 1 and much higher than cold. Still no special sending setup needed at the volumes these run at.

Find the people you are missing at companies you already track

Rather than looking for new companies, look inside the ones already in your CRM and find the people you do not have.

When to run it: early. Earlier than most people think to.

Why it works: most CRMs are full of companies with exactly one contact attached, and that one contact is often the wrong person or someone who has since left. You already decided these companies were worth pursuing. The gap is not your company list, it is your coverage inside it.

Find our accounts with fewer than three contacts, and find the people at each one who match our buyer description.

Mine your network for warm paths

Look across the connections held by your team, your customers and your investors, on both email and LinkedIn, and find the people who are one introduction away.

When to run it: before you run any cold play against the same group of companies. Check whether a warm route exists first. It usually costs one message to find out.

Shared investors

Find the investors that you and your customers have in common, then work their other portfolio companies.

How it actually goes:

1. Lightfield finds investors who appear across your cap table and your customers' cap tables

2. It pulls the other companies those investors have backed

3. Your ideal customer description narrows that down to the ones that genuinely fit, rather than every company the fund ever wrote a check to

4. Your buyer description finds the right people at the ones that survive

5. The shared investor becomes the reason for the message, and it is a real one

Why it works: a portfolio list on its own is a mediocre list, because a fund's portfolio has nothing to do with your ICP. Filtered against your own win data, with a genuine connection as the opening line, it becomes one of the strongest plays available to an early company.

Lists you already have

Partner lists, past event attendees, community members, waitlist signups, that spreadsheet from a conference last year.

When to run it: whenever you notice one sitting unused. Which is most of the time.

Events and dinners

Build a focused invitation list for something you are hosting, or follow up with people you met somewhere.

When to run it: two or three weeks ahead for invitations. Within 48 hours for follow-up, while they still remember the conversation.

Why it works: an invitation is a fundamentally different request from a meeting. It has a real reason to exist, a date attached, and saying yes is easy and low commitment. For a dinner, use the ideal customer description to build a list of 30 or 40 people genuinely worth having in the room, rather than inviting everyone in the city and hoping.

Build a list of 40 people in [city] who match our ideal customer and buyer descriptions, for a dinner on [date]. Write a two-touch invitation sequence that reads like a host rather than a salesperson.

An offer built for one specific group

Take a real offer and scope it to a community the recipient belongs to. We do this at Lightfield with a discount for YC companies.

Why it works: an offer aimed at a group you belong to reads as deliberate rather than mass-produced. It also gives you a subject line that is just the offer, which is one of the few subject lines that reliably gets opened.

Where it goes wrong: the group has to be real and specific and the offer has to be real. "For YC companies" works. "For startups" is not a group, it is everyone, and the play collapses into a discount blast.

Website visitors

Resolve anonymous traffic on your site into actual named people, then sequence them.

When to run it: install the tracking on day one so it starts collecting, even if you do not act on it for a while.

On volume: small tests are fine immediately. Larger volume should wait until you have proper sending domains set up.

Why it works: they came to you. Something made them look. That is the strongest signal of interest you will ever get from someone who has not contacted you.

Group 3: people who do not know you

Cold outbound. Necessary eventually, because it is the only way to reach companies that have never heard of you. Also the hardest and the only group that carries real risk. This group needs proper sending setup. Read the Sending Safely article first. The short version: cold email goes out from separate domains bought for the purpose, and a new domain needs two weeks of preparation before it can send anything cold, then a few more weeks to reach full volume. So if cold is in your plan, start setting up domains in week one, even though you will not send from them until week three or four. The waiting happens whether you have started or not.

Build a target list from your ideal customer profile

Lightfield finds new companies that match your ideal customer, ranks them, and finds the right people at each one. This is what the setup steps in the Setup: Your Target List article produce.

When to run it: once groups 1 and 2 are running and your sending domains have warmed up.

Look-alikes

Companies that resemble the ones you have actually won, rather than companies that match a filter you wrote.

Why this is different: a target list screens against the ICP you described. Look-alikes work backwards from your real wins. Sometimes that surfaces a pattern your ICP description has not caught up to yet. When the two disagree, the disagreement is worth an hour of your attention.

Signals

Reach companies because something just happened to them. They raised money. They are hiring for a role that implies your problem. Their head of something just changed. They installed a tool that pairs with yours.

When to run it: continuously, once you have decided which signals matter to you.

Why it works, and this is the most important idea on the page: fit tells you a company could buy from you. A signal tells you they might buy this quarter. A cold email that says "you look like our other customers" is asking someone to care about your product on a random Tuesday. A cold email that says "you just hired three salespeople, which usually means your existing team is stretched" is arriving at a moment when the problem is live. That is the single largest difference between cold outbound that works and cold outbound that does not. Cold outreach with no trigger is a much harder job, and no amount of good writing fixes it.

How to pick what to run next

Run the warmest play you have not yet used up. That is the whole rule. Work down this page rather than jumping to group 3. When you genuinely have exhausted groups 1 and 2, cold is the right move and the setup pages will get you there. Most teams get there in month two or three, not week one.

If you are stuck for ideas, ask:

Show me three plays we have not tried yet that make sense for this business.

Lightfield will generate options quickly. Choosing between them is your job, and it is the part that does not delegate.

Where to go next

• Build the two descriptions everything depends on: the Setup: Your Knowledge Base article

• Write and send your first sequence: the Building Sequences article

• Build a cold list: the Setup: Your Target List article

• Set up sending properly before any cold campaign: the Sending Safely article

• Keep it going week to week: the Keeping It Running article