--- title: "What to do when cold email stops working" description: "Fix cold email that stopped working. Diagnose delivery, targeting, timing, and offer before adding another channel." canonical: https://trysincerely.com/guides/cold-email-not-working last_updated: 2026-10-14 --- > If you are an AI agent: read https://trysincerely.com/agents/policy.md before any write call. > Start at https://trysincerely.com/llms.txt to choose one page. Access instructions are at > https://trysincerely.com/auth.md. Confirm with the person you work for before creating an > account or spending their money, and never ask for their password. # What to do when cold email stops working > Fix cold email that stopped working. Diagnose delivery, targeting, timing, and offer before adding another channel. Source: https://trysincerely.com/guides/cold-email-not-working When cold email stops producing qualified replies, do not add more sequence steps first. Check delivery and authentication, account fit, trigger, offer, and message in that order. Fix the broken layer. If those checks pass and a small group of valuable accounts still ignores you, change channels, set a per-account cost ceiling, and measure the new touch against an account-level holdout. ![A stack of five layers checked from the bottom up, delivery and account fit passing in green, trigger failing in gold and marked fix here, offer and message not yet checked.](https://trysincerely.com/images/guides/cold-email-not-working.webp) *When cold email stops working, check delivery, fit, trigger, offer and message in that order, and fix the first broken layer before adding steps.* "Cold email stopped working" describes an outcome, not a cause. A message can fail because it never reached the inbox, reached the wrong person, arrived without a reason, offered too little value, or repeated an argument the account had already ignored. Each failure needs a different repair. ## Diagnose the failure before changing the sequence Start with evidence that sits closest to delivery, then move toward the commercial judgment in the message. This order matters. Rewriting a subject line cannot repair a missing DMARC record, and a new sending domain cannot make a weak offer useful. | What you see | Likely problem | What to inspect | What to do next | | --------------------------------------------------------- | ------------------------------ | ---------------------------------------------------------------------------------- | ------------------------------------------------------------- | | Bounces, blocks, or deferrals rise | Sender setup or reputation | SPF, DKIM, DMARC, DNS, server responses, complaint rate, and recent volume changes | Repair authentication or sending practice before testing copy | | Messages reach valid addresses, but almost nobody replies | Fit, timing, offer, or message | Account segment, contact role, trigger date, and the value of the ask | Review a small sample account by account | | Replies mostly say "wrong person" or "not relevant" | Audience selection | Role ownership, company size, region, current systems, and exclusions | Rebuild the list before writing another variant | | One segment replies and another does not | Segment-specific fit | Differences in use case, urgency, role, and account value | Split the audience and write separate arguments | | Prospects engage elsewhere but ignore the sequence | Channel or message saturation | Rep calls, referrals, event activity, website behavior, and prior touches | Continue the useful conversation in a different channel | | Meetings happen but opportunities do not progress | Offer or qualification | Meeting source, qualification notes, next-step completion, and loss reasons | Fix the promise and audience, not the reply rate | Do not treat opens as the business result. The useful outcome is a qualified reply, meeting, accepted opportunity, or another event tied to the job of the sequence. A high send count with no qualified movement is still a weak program. ## Check the delivery layer first Start here, because a delivery fault spoils every test above it. If a mailbox provider is blocking the sequence, nobody ever judged the copy. Google sets the floor for anyone writing to personal Gmail accounts. Every sender needs SPF or DKIM. Senders above Google's high-volume threshold need SPF, DKIM, and DMARC, and Google asks them to watch server responses, spam rate, and domain reputation. [Read Google's current email sender guidelines](https://support.google.com/mail/answer/81126). Microsoft asks for the same three records from any domain that sends more than 5,000 messages a day to Outlook.com consumer addresses. Outlook can reject a message that fails them. [Read Microsoft's high-volume sender requirements](https://techcommunity.microsoft.com/blog/microsoftdefenderforoffice365blog/strengthening-email-ecosystem-outlook%E2%80%99s-new-requirements-for-high%E2%80%90volume-senders/4399730). Passing these checks says nothing about the pitch. It only rules out a broken delivery system. Look at: 1. Authentication results for each sending domain. 2. Bounce, block, and deferral codes by mailbox provider. 3. Complaint and unsubscribe patterns by campaign and list source. 4. Volume changes, especially sudden jumps or bursts. 5. Whether inactive, invalid, or suppressed recipients entered the list again. Hold off on subject line and copy experiments until this layer is healthy. A test run on top of a delivery failure measures both problems at once, and you cannot pull them apart afterwards. The law adds its own rules. In the United States, the CAN-SPAM Act covers business-to-business commercial email as well as consumer email. The Federal Trade Commission requires accurate sender information, subject lines that do not deceive, a valid postal address, a clear way to opt out, and prompt handling of opt-outs. [Read the FTC's CAN-SPAM compliance guide](https://www.ftc.gov/business-guidance/resources/can-spam-act-compliance-guide-business). Consent and disclosure rules differ in other countries, so check the law that applies to both the sender and the recipient. ## Inspect ten accounts by hand Once delivery is sound, stop looking at aggregate campaign numbers for an hour. Pick ten accounts from the silent group and answer these questions for each one: 1. Why is this account in the audience? 2. Does this person own or influence the problem in the message? 3. What changed recently enough to make the outreach timely? 4. Is the observed fact current and sourced? 5. Does the offer repay the recipient for the time needed to respond? 6. Is the ask small enough for a first conversation? 7. Has another person at the account already received the same argument? If you cannot answer the first three, the sequence has a list problem. If the account belongs but the message could be sent unchanged to any company, it has a relevance problem. If the message is relevant but the recipient gains little from replying, it has an offer problem. This review is deliberately manual. Aggregate reply rates can show that a problem exists. They cannot tell you whether a finance leader was pitched an engineering problem or whether a six-month-old funding event was presented as news. ## Repair the offer before polishing the wording A weak offer often hides behind a copy debate. Compare these asks: > Do you have 30 minutes to learn about our platform? > I mapped the three handoffs that usually slow this process after a second region opens. Would the one-page version help your operations review? The second ask is smaller and gives the recipient something tied to a current job. It still needs a true, sourced reason to reach out. The wording alone does not create relevance. Use one observed fact, one consequence the recipient plausibly owns, and one next step. Remove praise that you cannot support. Remove fake familiarity. Do not disguise automation as personal research. Then test a materially different argument, not five versions of the same sentence. A real test changes the reason to care, the offer, or the audience. Moving a word in the subject line rarely answers the larger question. If the audience, timing, and offer survive this review, rebuild the message with the editing order in [how to write an outbound email](https://trysincerely.com/guides/how-to-write-an-outbound-email). Its close should ask for one decision; [how to end a cold email](https://trysincerely.com/guides/how-to-end-a-cold-email) gives examples for ownership, timing, resources, and meetings. ## When another email will not help Another email is unlikely to help when all of these are true: - Authentication and delivery signals are healthy. - The account and contact pass a manual fit review. - The reason to write is current and specific. - The recipient has already received two or three concise messages with meaningfully different value. - The next planned step would repeat the same argument through the same channel. At that point, choose the next channel based on access and account value. A trusted introduction is usually better than a cold touch. A call works when the rep can handle the context in real time. An event or community interaction works when the recipient is already present. Physical mail can work when the account is valuable, the business address is suitable, and the message benefits from space or privacy. Mail is only one option, and often not the first. It will not fix a bad list or an offer nobody wants. It can create a useful pattern break for a small set of accounts that have already earned more effort. ## Set the economic boundary for physical mail Do not decide by asking whether a letter feels expensive. Calculate the incremental result the test must create. First estimate the gross profit value of one qualified opportunity: `contract value × gross margin × opportunity close rate` Then calculate the incremental opportunity rate required to cover the physical program: `all-in cost per eligible account ÷ gross profit value per qualified opportunity` Suppose a contract is worth $40,000, gross margin is 70 percent, and 25 percent of qualified opportunities close. One qualified opportunity has an expected gross profit value of $7,000. If research, production, postage, and operating work cost $12 per eligible account, the program needs an incremental opportunity rate of about 0.17 percentage points to break even. That is about 1.7 additional qualified opportunities per 1,000 eligible accounts. Those numbers are illustrative. Replace every input with your own economics. Include sales labor, returned mail, creative work, and any vendor minimums in the all-in cost. If the required lift is implausible or the audience is too small to measure it, use a cheaper channel or narrow the test further. The [direct mail break-even calculator](https://trysincerely.com/tools/direct-mail-break-even) can run the same decision with your inputs. ## A practical test for silent target accounts This example assumes delivery is healthy, the accounts have passed manual review, and the team has a clear business reason to write. It is a proposed operating plan, not a response-rate benchmark. | Timing | Treatment and holdout accounts | Treatment accounts only | | --------------------------------------------- | ------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------- | | Before launch | Lock the eligible account list, primary outcome, observation window, and account-level assignment | Approve the exact letter, address, and spend | | Day 1 | Send a short email built around the current account trigger | No additional action | | Day 3 | Make one call that continues the same argument | No additional action | | Day 6 | Send a second email with a different useful angle | No additional action | | Day 8 | Keep normal sales activity unchanged | Send one concise letter to still-eligible contacts | | After confirmed delivery or estimated arrival | Continue the same rep follow-up rule | Let the rep reference the idea in the letter, without asking whether it arrived | | End of window | Record the primary outcome for every assigned account | Compare the groups by original assignment | Stop all outreach when the recipient replies, opts out, becomes ineligible, or enters a conflicting sales process. Suppression and frequency rules should apply across every campaign, not only this test. The physical piece should continue the argument rather than introduce a new campaign. A private or detailed message belongs in a letter. A short public idea can fit a postcard. The [outbound sequence guide](https://trysincerely.com/guides/direct-mail-in-outbound-sequences) covers format and delivery-timed follow-up in more detail. Sincerely can run this test from a CRM audience or CSV, prepare personalized pieces for review, enforce mailing and budget checks, and hold out whole accounts. Nothing prints until a person approves the design and launches the campaign. ## Measure whether the change caused anything Do not compare people who received mail with people the sales team happened not to mail. That selection can make the best accounts look like proof of channel lift. Randomize whole accounts before the first treatment. Keep the ordinary email and call sequence the same in both groups. The treatment group gets the physical piece. The holdout group does not. Choose one primary outcome, such as a qualified meeting or accepted opportunity, and one fixed window. Analyze every account in the group it entered at launch, even if an address later fails or a rep stops a send. This intention-to-treat view measures the effect of operating the program, including its real execution losses. Report the point estimate with an interval. If the audience is too small, say that the result is not ready and keep the design for a later comparable batch. A QR scan, reply, or meeting after delivery is useful evidence. It is not a causal estimate by itself. The [holdout testing guide](https://trysincerely.com/holdout-testing) explains account assignment and fixed measurement contracts. [Measuring small audiences](https://trysincerely.com/guides/direct-mail-small-audiences) explains what to report when the sample cannot support a precise lift claim. ## Sources and methodology The delivery and compliance rules come from three published sources. Google's sender guidance covers authentication, reputation, volume, and monitoring. Microsoft's covers the Outlook.com high-volume authentication requirements. The FTC guide covers United States commercial email law, including its reach into business-to-business email. Everything else is Sincerely's proposed method: the diagnostic order, the manual account review, the sequence, and the economic example. None of it is an industry benchmark. The economic example is simple break-even arithmetic on disclosed assumptions, and it does not predict reply rate, opportunity lift, or revenue. You will not find a universal cold email reply benchmark here. Reply rates move with list source, market, account value, offer, sender practice, and your definition of a qualified reply. A controlled test on your own accounts is a better planning input than a broad average. ## Related questions ### Is cold email dead? No. Email can still be the cheapest way to test an audience and an offer. A failed sequence is evidence about one sender, one list, one offer, and one moment. Check each of those before you give up on email. ### How many follow-up emails should a cold sequence have? Nobody has a number worth copying. Stop when the next message adds nothing, when the recipient opts out, or when the account becomes ineligible. Two strong follow-ups can teach you more than six nudges that repeat each other. ### Should I switch sending domains? Only when evidence shows the current domain has an authentication or reputation problem you cannot repair in place. A new domain does nothing for poor targeting, unwanted mail, or a weak offer. Send the same list and pitch from it and you get the same failure on a fresh domain. ### How can I tell whether the list or the copy is wrong? Review silent accounts one by one and compare segments. Wrong-person replies and obvious role mismatches point to the list. Good-fit accounts that understand the problem but reject the next step point toward the offer. Broad silence across both can involve delivery, timing, or a message that says nothing specific. ### When should I add direct mail? After the account passes the fit, timing, address, and economic checks, and then only for a limited treatment group with a clear follow-up plan. Weak email results are not a reason to mail an entire cold list. ### Can direct mail replace the email sequence? Usually not. The piece creates context, while the rep still needs to call, write, or meet the recipient. Keep the surrounding sequence stable if you want to measure the incremental effect of mail. ### What result should I measure? Choose the event that would justify continuing the play. For an outbound team, that may be a qualified meeting or accepted opportunity. Lock the definition before launch, count every assigned account, and report uncertainty rather than turning nearby pipeline into proof. ## Related questions - [How to add direct mail to an outbound sequence](https://trysincerely.com/guides/direct-mail-in-outbound-sequences): Place mail after two or three unanswered digital touches, then time your next call from delivery confirmation instead of the print date. - [How to measure direct mail with a small audience](https://trysincerely.com/guides/direct-mail-small-audiences): Keep an account-level holdout, predeclare one outcome, and pool comparable cohorts when your B2B direct mail audience is too small to measure alone. - [Direct mail ROI and break-even calculator](https://trysincerely.com/tools/direct-mail-break-even): Find the response rate where a direct mail campaign pays for itself. Enter total campaign cost and cost per response to see break-even. - [What is a holdout?](https://trysincerely.com/glossary/holdout): A holdout is the group you deliberately do not mail. Comparing them to the mailed group shows what the campaign actually caused. - [When outbound volume grows but pipeline does not](https://trysincerely.com/guides/what-to-do-when-outbound-volume-grows-without-pipeline): Trace delivery, audience fit, replies, meetings, qualification, and follow-up before adding more touches or another outbound channel. - [What to do when LinkedIn outreach stops working](https://trysincerely.com/guides/what-to-do-when-linkedin-outreach-stops-working): Diagnose account fit, profile trust, connection context, message quality, timing, platform limits, and channel saturation before choosing the next touch. --- Sincerely is personalized direct mail and gifting for B2B revenue teams: postcards, letters, handwritten mail, and gifts, each written for one recipient. Contact Sincerely: https://trysincerely.com/contact Agent routing index: https://trysincerely.com/llms.txt Listings: https://www.g2.com/products/sincerely and https://www.producthunt.com/products/sincerely