A week is a method, not a close rate.
You can run find, prove, sample, and send in five working days. Nobody can print the probability the owner pays.
Guides
Find, prove, attach a DEMO, send. Useful even if you never buy Nerxus.
You can run find, prove, sample, and send in five working days. Nobody can print the probability the owner pays.
Search is a read. Mail is a write. Mixing those jobs on day one is how a new sender dies.
A specific finding plus a working preview beats a paragraph about your agency. If you cannot attach a live link, wait a day.
Pick one category a directory already uses and one city you can name on a map, then stop adding adjectives. Dentists in Austin is a hunt. Quality wellness brands in Texas is a paragraph you will retype all week.
The vertical has to be a job you can fulfill by Friday. If you build sites and booking blocks, hunt businesses with a missing site or a phone-only listing. If you only write Google posts, do not hunt ADA gaps. Week one dies when the offer and the detector belong to two different people.
The metro has to be small enough that you can read ten listings before lunch. A whole state is a vanity filter. A ZIP you cannot pronounce is a delay. If you already know three owners in one town, start there and still write the hunt as category plus city so you can repeat it next Monday without relying on memory.
Write the hunt in one line you could hand to an intern: category, city, the problem you sell, the thing you will not take. That line is an ICP. In Nerxus One it becomes a profile Scout can run again. In a notes app it is still the only sentence that keeps Thursday from turning into a new niche.
Do not pick two verticals to feel diversified. Two verticals is two fulfillment stacks, two sample types, and two ways to stall. One stack, one city, one problem. If that city is empty after an honest search, change the city, not the story of who you are.
Open the directories you already trust and search the category in that city. Google Maps and Google Business Profile are enough for week one. Facebook, Yelp, and Yellow Pages add coverage later. You are looking for a public problem, not a vibe.
A public problem is something you can screenshot without logging in as the owner. No website on the listing. A website that serves HTTP only. No booking button. A rating that sat at 4.6 and now reads 3.9. An unclaimed profile. A service the shop already does that never appears on the listing. Those are detectors. A nice storefront photo is not a detector.
Save the row the first time you decide it is real. Name, address, phone, listing URL, the problem in six words, the date you looked. Deduplicate on the listing’s own ID or you will email the same shop twice from two tabs. In Nerxus One, Scout does that dedupe and LeadScore ranks the cheap tags. In a sheet, you do it by hand and you still refuse to invent a score that means close rate.
Stop at a pile you can diagnose, not a pile that looks impressive. Twelve saved rows with one tag you can fulfill beats eighty names. Scout never contacts anyone, and neither should you from the search screen. If you keep a personal Gmail next to the map, you will write a generic pitch before you have a finding.
Importing a rented B2B file is optional and usually a downgrade this week. Those files are titles. You need listings with leaks. If you already paid for a list, score it for the same public problems and throw out the rows that are only a name.
Prove means you can show the owner the leak in their own public pages, with a sentence that names what you saw and when. It does not mean a twenty-page audit. An audit PDF tells them to hire someone else. A diagnostic is a memo: headline, evidence, what you would build, what you would charge.
Pick three rows from the pile, not the whole pile. Open the listing, the site if it exists, the booking path, the review count, the NAP consistency. Write what a stranger can see. If you cannot point at a URL or a screenshot, you do not have a finding yet. You have a hunch, and hunches become scan mail.
If you model a dollar, label it modeled and keep the assumption on the same line. Chair-hours lost to no-shows is an estimate that needs their book, their show rate, and their fee. You do not have those on Tuesday. You can say the listing has no online book button and that shops in this category take those bookings on the phone. You cannot say they are losing a number you invented.
Radar is the name of that memo in this product: bulk scan the list, deep-diagnose the one you would take, stamp a fulfillment action on the finding. Atlas can brand it. You can do the same work in a one-page note. The rule is the same. Evidence first. Dollars only with assumptions. No fake ROI to make the page look finished.
If two sources disagree, write the disagreement. Google says they have a site. The site 404s. That is a better sentence than a blended score. Unknown is a first-class state. A green check that came from nowhere is how you lose the meeting when the owner opens the laptop.
Build a preview that fixes the finding you already wrote down, and stop when a stranger can click it. If the leak is no website, the sample is a one-pager with their name, their city, their services, and a way to contact them. If the leak is no booking, the sample is a booking block on a page that uses their hours. If the leak is a collect-page gap, the sample is a page with public review buttons for every visitor.
DEMO is the product word for that preview. It is not a sales meeting. DEPLOY is a later, gated action. Week one only needs the DEMO. Host it on a URL you control. A localhost tab you cannot send is not a sample. A Figma file is not a sample. A Loom of you talking about a sample is not a sample.
Use their public facts and nothing you would have to guess. Name, category, city, phone from the listing, services they already print. Do not invent a testimonial. Do not invent a price list. Do not put a stock photo of a smiling family on a dental page and pretend you researched them. The owner will smell the generic pack in four seconds.
In Nerxus One, the finding already carries a fulfillment action id, so Forge opens the matching builder in the same sitting. One Free Fix as a magic button is not the thing to market if you cannot demo it. Market the DEMO asset. If you are outside the product, open the builder you already know and time-box two hours. If it is not clickable after two hours, the finding was too big for week one. Shrink the finding.
Keep a before and after. The before is their live listing or live URL. The after is your preview. That pair is the meeting. If you cannot put both links in one email, you are not ready to send.
Write the first sentence so it names the business, the public problem, and the live link. I noticed your Austin listing still has no website, so I built a one-page preview with your services and a book button. Here is the URL. That is the whole pitch. The rest of the email is one ask: fifteen minutes this week.
Do not open with your origin story. Do not open with AI. Do not open with a compliment about their community roots. Those lines are what a scan sounds like when it has no finding. The owner did not ask for a partner. They have a leak you can point at.
Stay inside facts they can check in thirty seconds. If you mention a rating, use the number you saw and the date. If you mention SSL, say the padlock is missing on the URL you visited. If you mention a missing service page, name the service they already advertise on a truck or a Facebook about section. Specific is polite. A dump of every detector you ran is a scan.
Relay writes from the finding in this product. Sequences have names like No-Website Wedge and Chair-Hour Loss. You can write the same letter in any editor. The rule is the finding first, the DEMO second, the ask third. One link. One meeting. No calendar maze, no six-page PDF attached beside the sample.
Read the draft out loud. If you would delete it as the owner, delete it as the sender. If the DEMO link 404s in your own browser, you do not have a letter yet.
Send from a domain you control, with a real From name, a physical address, and an unsubscribe that works. Unverified mail from a free inbox is how week one becomes a month of spam folders. In this product, your verified sending domain is required and unverified is a blocking gate. Outside the product, the same rule still holds.
Cap the first day. Five specific emails is a send. Fifty is a blast you cannot stand behind. Warm a new domain the boring way: a few real threads to people who know you, then the five. There is no honest shortcut that looks like a number we will print.
Quiet hours and stop-on-reply are not optional manners. If they answer, the sequence ends. If they unsubscribe, the sequence ends. If they say not now, you park the row and you do not invent a reason to keep poking. Pulse can hold that row on a signal-named board. A notes column that says parked on 3 September does the same job.
Do not add SMS. Do not add a cold call from a spoofed local number. Voice in this product is inbound only. SMS needs a consent checkbox that cannot arrive pre-checked. Week one is the wrong time to test whether those locks are real. They are real. They are also how you keep a client later.
ComplianceGate is the name of the chokepoint that brands the send. You can run a personal checklist that is uglier and still honest: domain verified, unsubscribe present, finding specific, DEMO live, volume sane, Supervised meaning you hit send yourself. Autonomous dispatch is not a week-one setting.
Classify the reply before you write back. Interested, question, not now, unsubscribe. Those four cover almost every inbox. Interested gets a time and the same two links. A question gets the shortest true answer. Not now gets a one-line thanks and a date you will look again. Unsubscribe gets silence.
The calendar in the mail is your booking link, not their patient book or their job calendar. Mixing those grids is how a marketer double-books a cleaning and a sales call. In Nerxus One those calendars stay isolated on purpose. In week one, use one Calendly or one Google booking page that is clearly yours.
On the call, walk the before and after. Do not walk your slide about the agency. Open their listing. Open your DEMO. Ask what is wrong with the preview. Fix the thing they name if it is small. Write the FEE as a sentence: I will put this live and connect the book button for this amount, this month. If you need a retainer to keep reviews or posts, say that as a second sentence, not a bundle that hides the first job.
Do not show a modeled dollar as if it were their P&L. If you used a chair-hour estimate in the memo, say the assumptions again and let them replace the numbers. Owners forgive a labeled estimate. They do not forgive a confident fake.
If they want a contract, write a one-page scope: the finding, the deliverable, the FEE, what deploy means, what you will not do. You are not their lawyer. You are the person who will push a page live. Steward is the later home of that record. A folder with the signed PDF is enough for client one.
Price the work you will finish, in a number you can invoice without a story. A one-pager with booking is a FEE. A collect page is a FEE. A month of review replies is a retainer you can describe as hours and artifacts, not as a growth multiple.
DEV and FEE are different dollars. DEV is a modeled monthly recovery for the SMB. FEE is what you charge once, or what you charge per month for ongoing work. Forecasts that mix them are a lie. Week one should almost always quote FEE. If you mention DEV, keep it in the memo with assumptions and keep it off the invoice.
Do not copy a price you saw on another site. This site will not invent a floor for Nerxus One until commercial policy exists, and you should not invent a floor for a dentist because a thread said agencies charge a round number. Look at your week. Look at the deliverable. Look at whether you can do the same job for the next three shops. Then say the number.
If they ask what it is worth, answer with the leak and the sample, not a multiplier. The listing has no site. Customers who search that category on Maps bounce to a competitor who has one. You built the missing page. That is the value sentence. A 3x ROI badge is a different product, and it is not this one.
Write down what you will not include. Logo design from scratch, photography, copy for ten service pages, a custom app. First clients expand because you are grateful. Grateful scope is how week one becomes an unpaid month.
The day they pay, the record has to become a client, not a leftover lead. Copy the finding, the DEMO URL, the deploy checklist, the billing email, and the owner’s preferred contact into one place you will still open in ninety days. In this product, closing hands the same business uuid to Steward. There is no second import. In a sheet, that means you stop using the prospect tab as the client tab.
Give them a view that is not your notes. A portal, a shared folder, or a weekly email they can forward. Plain language. No internal comments. No costs you charge other people. No health score that prints their revenue. Health here is Healthy, Watch, or At-risk. That rule exists so a screenshot in a group chat does not become a leak.
Fulfill what you sold before you hunt client two. Deploy is gated for a reason. Preview and production are different buttons. If you promised a book button, the button has to land on a number or a calendar that belongs to them. If you promised a collect page, public review buttons render for every visitor. Private recovery can sit beside that page. It cannot replace it.
Park leftover findings as upsells, not as surprise invoices. Radar often leaves more than you sold. That is a gift if you wait. It is a bait-and-switch if you spring a privacy-policy draft on a site job they already paid for, unless the SOW said so.
Set the next date before you hang up. Review replies on Fridays, a Watchtower-style re-check of the listing in thirty days, a renewal conversation before the card expires. Client one dies in month two when the marketer goes back to Maps and treats the paying shop as done.
Put a category a directory already uses, a city you can point at, the problem you sell, and the thing you will not take in the ICP line before you open a map. Dentists in Austin with no online booking, no DSOs is a hunt. Quality wellness brands in Texas is a paragraph you will retype all week. The line is an ICP. In Nerxus One it becomes a profile Scout can run again. In a notes app it is still the only sentence that keeps Thursday from turning into a new niche.
The vertical has to be a job you can fulfill by Friday. If you build sites and booking blocks, hunt businesses with a missing site or a phone-only listing. If you only write Google posts, do not hunt ADA gaps. Week one dies when the offer and the detector belong to two different people. Do not pick two verticals to feel diversified. Two verticals is two fulfillment stacks.
The metro has to be small enough that you can read ten listings before lunch. A whole state is a vanity filter. If you already know three owners in one town, start there and still write the hunt as category plus city so you can repeat it next Monday without relying on memory. If that city is empty after an honest search, change the city, not the story of who you are.
You do not need Nerxus One to write the line. You need the four fields. When you later have a seat, Scout, Radar, Forge, Relay, and Steward are how those fields become a uuid, a memo, a DEMO, a gated send, and a client record. The method does not wait on the waitlist. It waits on a finding, a live sample, and a send you can defend.
Host the DEMO on a URL you control that a stranger can open on a phone, because a localhost tab you cannot send is not a sample. A Figma file is not a sample. A Loom of you talking about a sample is not a sample. In this product, Forge opens from the fulfillment action and the preview is the asset. Outside it, open the builder you already know and time-box two hours.
Use their public facts and nothing you would have to guess. Name, category, city, phone from the listing, services they already print. Do not invent a testimonial. Do not invent a price list. Do not put a stock photo of a smiling family on a dental page and pretend you researched them. The owner will smell the generic pack in four seconds.
Label the preview. Draft preview. Not live. Nothing changes until you say so. That sentence lowers the threat. DEPLOY stays gated in this product for the same reason. Copy the habit. If the finding is no website, the sample is a one-pager with a way to contact them. If the leak is no booking, the sample is a booking block on a page that uses their hours. If the leak is a collect-page gap, public review buttons render for every visitor.
Keep a before and after. The before is their live listing or live URL. The after is your preview. That pair is the meeting. If you cannot put both links in one email, you are not ready to send. One Free Fix as a magic button is not the thing to market if you cannot demo it. Market the DEMO asset. If it is not clickable after two hours, the finding was too big for week one. Shrink the finding.
Write down leftover findings on the same record even if they do not buy, because Radar often leaves more than you sold and those leftovers are month two or a reason to come back, not a dump you mailed on Wednesday. A shop that said not now on a site DEMO may still have a collect-page gap. Park the row with a date. Pulse can hold that. A notes column that says parked on 3 September does the same job.
Do not spring a privacy-policy draft on a site job they already paid for unless the SOW said so. Do not spring leftovers on a shop that said no. Leftovers are a gift if you wait. They are a bait-and-switch if you invoice them on Friday. Health later is Healthy, Watch, or At-risk. A client's revenue never appears. Copy that habit in a sheet if you have no Steward yet.
If nobody replies, leftover findings are not the next metro. Re-read the three emails you sent. If the finding is vague or the link is dead, fix those two things before you hunt a new niche. Watchtower exists to re-scan what you saved and tell you what changed. If you are in a sheet, reopen the URLs. Do not trust last week's tags. Freshness decays.
When they do pay, leftover findings travel with the uuid. Steward is the later home of that list. A folder with the signed PDF and a leftover list they already saw is enough for client one. The rule is they saw the list. You can run this with a notes file. Nerxus One is the version where the finding, the DEMO, the letter, and the client record share an id.
Week one looks like five specific emails on a domain you already control if the domain is verified, because waiting on DNS is the feature when you do not have one and speed is the temptation when you do. Unverified mail from a free inbox is how week one becomes a month of spam folders. In this product, your verified sending domain is required and unverified is a blocking gate. Outside the product, the same rule still holds.
Cap the first day even if the domain is old. Five specific emails is a send. Fifty is a blast you cannot stand behind. Warm a new domain the boring way: a few real threads to people who know you, then the five. There is no honest shortcut that looks like a number we will print. Quiet hours and stop-on-reply are not optional manners.
Do not add SMS. Do not add a cold call from a spoofed local number. Voice in this product is inbound only. SMS needs a consent checkbox that cannot arrive pre-checked. Week one is the wrong time to test whether those locks are real. They are real. They are also how you keep a client later. ComplianceGate is the name of the chokepoint that brands the send. You can run a personal checklist that is uglier and still honest.
If the domain is already warm from another motion, you still do not paste a generic 12-step. The finding first, the DEMO second, the ask third. One link. One meeting. Autonomous dispatch is not a week-one setting. Supervised means you hit send yourself. Get started on this site is a waitlist until access opens in batches. The method does not wait on that door.
Tell a partner who wants to blast the metro that you will take three to seven shops you can fulfill, each with a finding and a sample, and that twenty generic emails is a worse week than five specific ones. If they resent that sentence, they want Instantly's sport. Instantly wins volume infrastructure. This guide is sample-first. Mixing those jobs is how a new sender dies.
Share the ICP as fields, not as a vibe channel. Two people running dentists / Austin / no booking / booking DEMO will produce comparable piles. Two people running quality Austin clinics will produce an argument. Do not contact anyone during the test search. The test is a read. Contact is a later module with a gate.
If they want to split fulfillment, write who builds the DEMO and who hits send. Two clickers on the same letter is how a sequence double-sends. If they want to share a From line across two firms, refuse. Isolation starts at client one.
You do not need to believe an overnight playbook already ran in production. Governance can be live while autonomous dispatch is still the open half. Supervised is how a careful marketer spends the first month. A partner who wants zero approval clicks came to the wrong default. Send them to a volume shop and keep the week.
Monday of week two runs the same hunt, then diffs what changed, because a new niche is how week one failures hide. Same vertical, same metro, or the next metro if the first one is thin. Watchtower exists to re-scan what you saved: a rating drop, a permit, a license drift, a site that appeared. Freshness decays. A stale saved lead is a polite way to waste a send. If you are in a sheet, reopen the URLs. Do not trust last week's tags.
Week one usually fails because the hunt was a mood, the proof was a PDF, the sample was a promise, or the send was a blast. Those four failures map to the four days. Fix the day that broke. Do not add a fifth tool. Do not add a second vertical to feel diversified. The other common failure is contacting from the search screen. If that was your Wednesday, Thursday is a rebuild, not a bigger list.
A third failure is selling a detector you cannot fulfill. ADA copy without a reviewer. A modeled chair-hour number without their book. A Yelp reply you cannot post because that network is monitor or assisted. Sell the finding you can finish. The rest waits. Re-rank your sellable list after ten conversations, not after a new detector ships.
The method in this guide does not wait on a seat. It waits on a finding, a live sample, and a send you can defend. Get started if you want into the next batch. The marketing site is not the product.
Week one usually fails because the hunt was a mood, the proof was a PDF, the sample was a promise, or the send was a blast. Those four failures map to the four days. Fix the day that broke. Do not add a fifth tool.
The other common failure is contacting from the search screen. You found a dentist with no booking, you felt lucky, you wrote from Gmail before you built anything. Sample-first sounded good in a demo and then you became every other vendor. If that was your Wednesday, Thursday is a rebuild, not a bigger list.
A third failure is selling a detector you cannot fulfill. ADA copy without a reviewer. A modeled chair-hour number without their book. A Yelp reply you cannot post because that network is monitor or assisted. Sell the finding you can finish. The rest waits.
Monday of week two, run the same hunt. Same vertical, same metro, or the next metro if the first one is thin. Watchtower exists to re-scan what you saved and tell you what changed: a rating drop, a permit, a license drift, a site that appeared. Freshness decays. A stale saved lead is a polite way to waste a send. If you are in a sheet, reopen the URLs. Do not trust last week’s tags.
You do not need to believe an overnight playbook already ran in production. Governance can be live while autonomous dispatch is still the open half. Supervised is how a careful marketer spends the first month. Get started on this site is a waitlist until access opens in batches. The method in this guide does not wait on that door. It waits on a finding, a live sample, and a send you can defend.
You can run the full loop in a week. A signed job is a separate outcome. Treat the week as a cadence, not a promise.
No. A Maps tab, a notes file, a preview you can host, and a sending domain are enough. The product names are how the same steps look in our app.
No. Week one is email-first on a domain you control. Cold voice and cold SMS are how new marketers get blocked and how this product stays locked.
Three to seven you can actually fulfill, each with a finding and a sample. Twenty generic emails is a worse week than five specific ones.
Charge a FEE you can say out loud for the work you will do this month. Do not quote a modeled recovery number as if it were cash in their register.
Re-read the three emails you sent. If the finding is vague or the link is dead, fix those two things before you hunt a new metro.
No. Name the leak and the live link. Modeled dollars belong in a memo with assumptions, not in line one of a cold email.
No. Week one is email-first on a domain you control. Cold voice and cold SMS are how new marketers get blocked.
A Maps tab, a notes file, a preview you can host, and a sending domain are enough. The product names are how the same steps look in our app.
Write them down on the record as month two. Do not spring them as a surprise invoice on Friday.
No. You can run the loop in a week. A signed job is a separate outcome. Treat the week as a cadence, not a promise.
Tell us your vertical, your metro, and what you can fulfill. Supervised is the default. Plans start at $99/mo.
14-day trial. Card required, converts on day 15. Full table on /pricing.