A lead generation agency red flag is any sign that the agency is selling you a number it cannot explain.
The common ones: a guaranteed count of meetings before anyone has studied your offer, no written definition of a qualified lead, lists nobody verifies, sending domains registered in the agency's name, a long term with no way out, and reports full of activity with no booked calls, held calls or signed deals in them.
Listing the flags is the easy part. Timing is where the money goes. A guarantee you question on the sales call costs you 10 minutes. The same guarantee discovered in month 3 costs a quarter's fees and a market that has already heard from you once.
So I have sorted the flags by where they surface: the sales call, the proposal, the contract, the first 30 days and the reporting after that. Each one comes with the question to ask and what to do when the answer is wrong. It works the same whether the firm calls itself a lead gen, outbound or appointment setting agency.
Smirnov Consulting Group is a Prague-based B2B outbound lead generation agency that runs cold email and LinkedIn campaigns for founder-led B2B companies and books qualified sales calls. So this is my own industry, and every question below is one you can put to us too.
What does missing a red flag actually cost?
There are two costs, and founders usually count only the first.
The first is fees. Take an agreement that runs 12 months with no exit before the end. Say the warning signs show up in month 2 and you read them correctly. You still pay 10 more months for a campaign you already know is wrong.
Now put the same discovery inside a 3-month first term. You pay 1 more month and leave. 10 months minus 1 month = 9 months of fees, and that gap is the price of one clause you did or did not read. Multiply it by your own monthly fee before you sign anything.
| When you spot the flag | What you can still do | What it costs you |
|---|---|---|
| Sales call | Ask one more question, or end the call | 10 minutes |
| Proposal | Ask for a rewrite, or walk | An afternoon |
| Contract | Change the clause before you sign | One awkward negotiation |
| First 30 days | Use the pilot or exit terms, if you have them | A month of fees plus the setup |
| Month 3 or later | Demand a fix plan, renegotiate or leave | Months of fees, a list you have burned, and time to rebuild |
The second cost never shows up on an invoice. One US client had 3 failed agencies behind them and a settled belief that cold email did not work in their niche, until a rebuild proved otherwise. I broke that campaign down in the post about lists deciding more than copy.
The fees of 3 agencies hurt. The conclusion they left behind nearly cost that company the channel that later worked for it.
Which red flags show up on the first sales call?
The sales call is the cheapest place to catch anything, because you have not agreed to a single thing yet.
A number before any questions. If you hear "40 meetings a month" before the agency has asked what you sell, who buys it and how long your sales cycle runs, that number was written before you joined the call.
Ask one question back: "What would have to be true on our side for that number to happen?" A good answer names conditions. Your offer, your proof, a list big enough, someone free to take the calls. A weak answer repeats the number with more confidence.
The agency controls the list, the copy and the sending. It does not control whether your market wants the offer, whether your team shows up prepared, or whether your calendar has room. A promise that ignores half the inputs is a sales line.
No questions about you. A good first call feels a little like being interviewed. Which clients do you want more of? What did you close last quarter? Who takes the booked calls? What have you tried before, and what happened?
If the whole call is their deck, they have no idea whether they can deliver for you, and they have chosen not to find out.
Nobody you can call. Logos and testimonials, but no client who will get on the phone. Ask for 2 references with a deal size close to yours, and ask those references what the agency got wrong.
I laid out a full method for testing whether an agency's case studies hold up. The short version: look for a result with a date range attached, and a client willing to confirm it out loud.
If 2 of these 3 show up on one call, do not book the second call.
What should worry you in the written proposal?
A proposal is where the promises from the call either turn into definitions or stay vague, now on paper. Read it for what is missing. If you have two on the desk, score both proposals line by line before the monthly fee gets a vote.
"Qualified" is never defined. Ask for the definition in writing: job titles, company size, the problem the person has, and whether they agreed to a call at a set time. Without it, every meeting number in the proposal is impossible to measure, and any dispute later becomes your word against theirs.
Pay per lead or per meeting, with no quality bar. Paying per result sounds safe. It is safe only when the result is defined and you can reject the ones that miss.
Look for a written rejection window: how many days you have to flag a meeting that did not fit, and what happens next. No window means you pay for whatever lands in the calendar.
One package for everyone. Replace your company name with a competitor's and read the proposal again. If it still reads fine, nobody planned your campaign. A real proposal names your buyer titles, the segments it would test first, and the reason for that order.
Silence about the list. The list decides more than the copy, so the proposal should say where the data comes from, how it gets checked and what bounce rate the agency accepts. My line is 0.5%. Every address gets verified with NeverBounce or DeBounce before a single email goes out to it, and the bounce rate stays below that line.
If an agency shrugs and calls bounces of a few percent fine, it is planning to spend your domain's reputation on its own shortcuts.
A monthly fee with the real costs outside it. Data, tools, sending domains, mailboxes, setup, CRM work. Some agencies fold these into the fee and some bill them on top. Either is fine if it is written down. Ask for the all-in cost of month 1 and of month 4, and compare those two numbers between proposals, not the headline fee.
Which contract clauses should stop you from signing?
The contract is the last stage where fixing a flag costs you nothing but a conversation. After the signature, every change is a negotiation you start from the weaker side.
Your domains and data sit in the agency's name. Sending domains, mailboxes, the lead list, the reply history, the CRM records. If the contract does not say which of these come back to you when you leave, assume none of them do.
Settle it in writing before the agency buys a single domain, including who holds the logins for the domain registrar and the mailboxes.
A long lock-in with no off-ramp. A 6- or 12-month term is not a red flag on its own, because outbound needs time. The flag is a long term with no pilot, no exit tied to results, and a notice period that quietly adds months.
The 9-month gap in the cost section above comes from exactly this clause. Published norms for minimum terms, notice and exit fees sit side by side in this comparison of lead gen agency contract terms.
A guarantee with soft edges. If the contract promises a number, read the definitions under it: what counts, who decides whether it counted, and what you get if it is missed. When the remedy is more months of the same service, the guarantee mainly keeps you in the contract.
No right to see the work. You should be able to see the list, the sequences and the reply inbox, or get a live export on request. A contract that keeps all of it behind the agency's dashboard leaves you judging the campaign only through the agency's own reports.
This is not legal advice. Before you sign, get the ownership and exit wording checked by your own lawyer.
What goes wrong in the first 30 days?
Month 1 is mostly setup, so do not judge it on meetings. Judge it on what the agency asks you for and what it shows you.
Sending starts before anyone asked you anything. No onboarding session about your offer, your best clients, your buyer titles and the objections your sales team hears every week. No request for case studies. No look at the LinkedIn profiles your prospects will open.
An agency that can launch without any of that runs one campaign for everybody.
The copy could belong to any company. In a 2023 post about hired SDRs and BDRs who charge a monthly fee upfront and cannot book meetings, I wrote that they "will use completely outdated strategies that worked 3 years ago and deliver nothing." Agencies can fall into the same habit.
You can spot a stale playbook in the first sequence sent to you for approval. A template with your name dropped in. One sequence for every title and industry. A "quick question" opener about a problem nobody on your side confirmed. Your prospects get that same version of you, and it is the one they remember.
No list sample before launch. Ask to see a slice of the real list before the first send, with a reason next to each person for why they are on it. Check a handful yourself. Wrong titles or companies that could never buy from you mean the campaign is broken before a single email leaves.
Setup you cannot check. Under the sender requirements Google publishes for Gmail (as of September 2026), every sender must authenticate its domain with SPF or DKIM. Bulk senders, which Google counts as more than 5,000 Gmail messages per day, need all three: SPF, DKIM and DMARC.
Ask the agency to show you those records on your sending domains. Whoever set them up can do it in a screen share. Delays or a vague answer are the flag.
If your agreement has a pilot or an exit at day 30, this is the month to write your observations down. You may need them.
How do you know a lead gen agency is failing after month 2?
From month 2 on you should be judging conversations, and by the end of month 3, calls that actually happened. The report has to let you do that.
The report is all activity. Emails sent, connection requests, open rates and "touches." Those rows prove the agency is busy. The rows that matter sit lower in the funnel: positive replies, booked calls, calls that took place, and deals that moved forward. If they are missing or blank, ask for them by name.
No deliverability numbers. Bounce rate and spam complaints belong on every report. For complaints, the same Google page sets 0.3% as a spam rate your domain should never hit in Postmaster Tools, and asks senders to stay under 0.1%. An agency that cannot tell you where your domains sit against those lines is not watching them.
Meetings that do not fit. People who took the call to be polite. Companies too small to buy. The wrong title, or no budget. One or two can slip into a good campaign too. A pattern is the flag.
Before the next review, go through every held call and mark each one yes or no against the written definition of qualified. The share marked no is a number the agency has to answer. There is a fuller method for scoring meetings an agency already delivered if the pattern is not obvious.
Numbers once a month, as slides. A monthly deck gives you 12 chances a year to catch a problem. Weekly numbers give you 52. Ask for the funnel every week, in a sheet you can sort and filter yourself.
Every fix is more volume. When results stall, the only answer on offer is to send more. More sends on a list and message that are not working buy you the same result faster, and burn more of your market on the way. Ask what they would change apart from volume.
They agree to everything. A new segment every week. A second offer squeezed into the same sequence. A campaign paused for 2 weeks because you got busy, then restarted cold. An agency that never pushes back is working to keep the contract, whatever it does to your pipeline.
What to do: ask for a written fix plan with 2 or 3 specific changes, the number each one should move, and a date. If the next month does not move those numbers, you have your answer and, with a decent contract, your exit.
What does a good agency do that a bad one does not?
Some signals point the other way. How an agency picks its clients and treats work tells you more than its pitch does.
It turns clients down. My own filter runs before any contract. I check the prospect's reputation and reviews.
The client has to have time to actually take the sales calls. I say no to companies that would pause the campaign the moment they get busy, to beginners without a proven offer, and to anyone asking us to build their business from zero.
An agency with no filter of its own takes your money whether it can help or not, and you find out about the mismatch in month 3. So ask every agency you talk to who they turned down recently and why. A real answer has a reason in it.
It pays for work. When I hire, I never ask people to complete free projects. Any work should be paid. Look for both sides of that rule when you buy. If an agency's own job ads ask SDR candidates for an unpaid test campaign, that shows how it values the people who will write to your prospects.
If an agency offers you weeks of free sending to win the deal, ask why its calendar has room for that. A short paid pilot with written definitions is the honest version of the same idea.
It fixes the foundations before it sends. The rebuild after those 3 agencies began with the offer and the buyer profile, and sending came last. An agency whose first week is all about launch dates has already told you where it will start.
A checklist to take into your next agency call
Here are the key questions from each stage in one table, with the answer that should end the conversation.
| Stage | Ask this | Walk away if you hear |
|---|---|---|
| Sales call | "What has to be true on our side for that number to happen?" | The number again, with no conditions |
| Sales call | "Can I speak to 2 clients with a deal size like mine?" | Logos and testimonials only |
| Proposal | "Define a qualified meeting in writing." | "You'll know one when you see it" |
| Proposal | "Where does the data come from, and what bounce rate do you accept?" | "Bounces of a few percent are fine" |
| Proposal | "What do month 1 and month 4 cost, all in?" | One number, with extras "depending" |
| Contract | "What do we keep if we leave?" | "We'll sort that out later" |
| Contract | "How do we exit if results stall?" | Only the end date |
| First 30 days | "Show me a list sample and the SPF, DKIM and DMARC records." | Delays, or a tool name instead of rows |
| Month 2 onward | "Where are held calls and deals on the report?" | Opens and activity only |
| Any time | "Who did you turn down recently, and why?" | Nobody |
Take the table into the call and mark each row as you go. The flags you catch there cost you minutes. The same flags in month 3 cost you a quarter.
What founders ask about agency warning signs
How many red flags are too many?
One is enough if it sits in the contract: unclear domain ownership, no exit, or no written definition of a qualified meeting. Fix those before signing or walk. On a sales call or in a proposal, 2 or more flags from the same stage usually mean nobody has thought about your campaign yet, and a second meeting rarely changes that.
Can a red flag be fixed mid-contract?
Some can. Reporting, list quality and copy can change within a month if the agency agrees to a written fix plan with numbers and a date. Ownership and exit terms are much harder to change after signing, because most of your bargaining power ends at the signature.
Is a written guarantee ever legitimate?
A guarantee is only as good as its definitions. If it defines a qualified meeting precisely, lets you reject meetings that miss, and pays a real remedy, you can judge it. If the remedy is more months of service, it mostly keeps you in the contract. Either way, ask what the agency needs from you for the number to happen.
I already signed and I see these flags. What now?
Write down each flag with a date and an example, then ask for a fix plan in writing: the changes, the numbers they should move, and by when. Check your contract for the notice period and what you keep on exit. If the next month moves nothing, give notice on the date that costs you least. This is not legal advice.