Why cold email needs a funnel, not a feeling
Cold email is one of the oldest ways to start a business conversation: you write to someone at another company who has never heard of you, explain in a few lines why you are getting in touch, and ask whether a short call makes sense. Done well, it is a polite, low-cost way to reach exactly the people who might need what you sell. Done badly, it is noise that annoys recipients, damages the sender's reputation and can break the law.
The difference between the two is rarely visible in a single email. It shows up in the numbers. Teams that run outreach by feeling tend to celebrate the wrong things: a high open rate, a busy inbox, a large number of emails sent. Teams that run it by the numbers ask a stricter question: out of every thousand people we wrote to, how many wanted to talk to us, and what did we have to do to the other nine hundred and ninety to get there?
This free lesson gives you the basic measuring kit for that question. You will learn the stages of the cold email funnel, the four rates that describe it and the one number that sums it up, meetings booked per 1,000 contacts. The example is Brackenmoor Labs, a fictional software company that sells shift-planning software to warehouses and logistics firms. All companies, people, campaigns and numbers are invented for learning, and every benchmark mentioned is illustrative.
The six stages of the funnel
Every cold email campaign passes through the same stages. Each stage is a count, and each count is smaller than the one before.
- Contacts emailed. The number of people who were put into the sequence. One contact is one person, however many follow-up emails that person later receives.
- Delivered. Contacts minus bounces. A bounce is an email that the receiving server rejected, usually because the address no longer exists.
- Opened. Contacts whose mail program loaded the email at least once. This is the least reliable count of all, as explained below.
- Replied. Contacts who wrote back, whatever they wrote. An automatic out-of-office note is normally not counted.
- Positive replies. Replies that show interest ("send me more", "let's talk") or point to the right colleague. "Not interested" and "please remove me" are replies, but not positive ones.
- Meetings booked. Positive replies that led to a date in the calendar.
Figure 1 shows the whole first quarter of Brackenmoor Labs in one picture. Of 10,250 contacts, 9,776 received the email, 431 replied, 155 replied positively and 82 booked a meeting. The steep drop is normal for cold outreach. It is also the reason why small differences in the rates matter so much: when only about four in a hundred people answer, one more answer per hundred is a large improvement.
Four rates and one result
Counts cannot be compared between campaigns of different size. A campaign with 2,000 contacts will almost always produce more replies than one with 500, even if it is much worse. Rates make campaigns comparable. For the cold email funnel you need these five formulas.
- Delivery rate = (contacts − bounced) ÷ contacts. The mirror image is the bounce rate, bounced ÷ contacts.
- Open rate = opened ÷ delivered.
- Reply rate = replied ÷ delivered.
- Positive-reply rate = positive replies ÷ delivered.
- Meetings booked per 1,000 contacts = meetings ÷ contacts × 1,000.
Notice the two different bases. The three response rates divide by delivered emails, because a person who never received the email had no chance to open or answer it. Measuring them on all contacts would mix two separate problems, a bad list and a weak message, into one number. The business result, on the other hand, divides by all contacts: every name on the list was paid for, researched and used up, whether the email arrived or not.
A fully worked example: Brackenmoor Labs, first quarter 2027
Brackenmoor Labs ran ten campaigns in the first quarter. Eight went to carefully built lists for one industry and one job title each. One went to an old badge list from a 2022 trade fair that nobody had verified, and one went to a small list of companies that resemble existing customers who had recommended the product. These are the raw counts, the same ones as in the lab above.
| Campaign | Contacts | Bounced | Opened | Replied | Positive | Meetings |
|---|---|---|---|---|---|---|
| 3PL operators · operations directors | 1,200 | 36 | 536 | 70 | 26 | 14 |
| Cold-chain logistics · site managers | 800 | 18 | 391 | 55 | 23 | 13 |
| E-commerce fulfilment · COOs | 1,500 | 82 | 610 | 57 | 17 | 8 |
| Parcel hubs · shift leads | 950 | 21 | 420 | 39 | 13 | 6 |
| Food wholesalers · logistics heads | 700 | 49 | 280 | 20 | 5 | 2 |
| Pharma distribution · operations | 600 | 12 | 300 | 41 | 19 | 11 |
| Furniture retail DCs · planners | 1,100 | 30 | 449 | 43 | 12 | 6 |
| Automotive parts · plant logistics | 900 | 27 | 367 | 35 | 14 | 8 |
| Trade-fair badge list 2022 (unverified) | 2,000 | 190 | 543 | 27 | 4 | 1 |
| Referral look-alikes · operations directors | 500 | 9 | 265 | 44 | 22 | 13 |
| Total | 10,250 | 474 | 4,161 | 431 | 155 | 82 |
Looking at the raw counts alone, the 3PL campaign seems to be the star with 70 replies and 14 meetings, and the trade-fair list looks respectable with 543 opens, more than most campaigns. The rates tell a different story.
| Campaign | Delivery rate | Open rate | Reply rate | Positive-reply rate | Meetings per 1,000 contacts |
|---|---|---|---|---|---|
| 3PL operators · operations directors | 97.0% | 46.0% | 6.01% | 2.23% | 11.7 |
| Cold-chain logistics · site managers | 97.8% | 50.0% | 7.03% | 2.94% | 16.3 |
| E-commerce fulfilment · COOs | 94.5% | 43.0% | 4.02% | 1.20% | 5.3 |
| Parcel hubs · shift leads | 97.8% | 45.2% | 4.20% | 1.40% | 6.3 |
| Food wholesalers · logistics heads | 93.0% | 43.0% | 3.07% | 0.77% | 2.9 |
| Pharma distribution · operations | 98.0% | 51.0% | 6.97% | 3.23% | 18.3 |
| Furniture retail DCs · planners | 97.3% | 42.0% | 4.02% | 1.12% | 5.5 |
| Automotive parts · plant logistics | 97.0% | 42.0% | 4.01% | 1.60% | 8.9 |
| Trade-fair badge list 2022 (unverified) | 90.5% | 30.0% | 1.49% | 0.22% | 0.5 |
| Referral look-alikes · operations directors | 98.2% | 54.0% | 8.96% | 4.48% | 26.0 |
| All campaigns | 95.4% | 42.6% | 4.41% | 1.59% | 8.0 |
The formulas step by step
In the lab the campaign names are in column A, the six counts in columns B to G, and the first campaign is in row 4. All five formulas are written once in row 4 and copied down to row 13.
- Delivery rate (H4):
=(B4-C4)/B4. For the 3PL campaign: (1,200 − 36) ÷ 1,200 = 97.0%. The brackets matter. Without them Excel would divide C4 by B4 first and subtract the result from B4. - Open rate (I4):
=D4/(B4-C4). 536 ÷ 1,164 = 46.0%. Here the brackets belong in the divisor: the delivered emails are B4 − C4. - Reply rate (J4):
=E4/(B4-C4). 70 ÷ 1,164 = 6.01%. - Positive-reply rate (K4):
=F4/(B4-C4). 26 ÷ 1,164 = 2.23%. - Meetings per 1,000 contacts (L4):
=G4/B4*1000. 14 ÷ 1,200 × 1,000 = 11.7. This one divides by the contacts in B4, not by the delivered emails. - Select H4:L4 and drag the fill handle down to row 13, or double-click it. Because all references are relative, each row uses its own counts. Format H to K as percentages and L with one decimal place.
=B4-C4 and divide by that column instead. The result is the same, the formulas are shorter, and the delivered count is useful on its own. The lab sheet has no spare column, so there you write B4 − C4 inside each formula.How to read the result
Start with the delivery rate, because it tells you whether the other numbers can be trusted at all. Eight campaigns deliver between 94.5% and 98.2%. The food wholesaler list is weaker at 93.0%, and the old trade-fair list is in a class of its own: 190 of 2,000 emails bounced, a delivery rate of only 90.5%. A list like that should never have been sent. Lesson 2 of the course shows how verification catches this before the first email goes out.
Next comes the reply rate. The referral look-alikes lead with 8.96%, followed by cold-chain logistics with 7.03% and pharma distribution with 6.97%. The trade-fair list reaches 1.49%. Its 543 opens looked like interest, and its replies show that there was hardly any, which is one more reason not to steer by opens.
The positive-reply rate sharpens the picture, because it removes the answers that are really refusals. In the referral campaign half of all replies are positive (22 of 44). In the trade-fair campaign only 4 of 27 are, about 15%; most of the people who answered wrote to say no. Across all campaigns 36% of the replies are positive.
Finally, the result: meetings per 1,000 contacts. The referral look-alikes produce 26.0, pharma distribution 18.3 and cold-chain logistics 16.3. The 3PL campaign, the apparent star of the raw counts, is good but not the best at 11.7. The trade-fair list produces 0.5: two thousand names for a single meeting.
Totals: add the counts, never average the rates
What is the positive-reply rate of the whole quarter? The tempting answer is to average the ten campaign rates, which gives 1.92%. That is wrong, because it gives the small referral campaign with 500 contacts the same weight as the trade-fair list with 2,000. The correct way is to add up the counts first and divide the sums: 155 positive replies ÷ 9,776 delivered emails = 1.59%.
The same applies to every rate in the bottom row of the result table. The overall reply rate is 431 ÷ 9,776 = 4.41%, and the quarter produced 82 ÷ 10,250 × 1,000 = 8.0 meetings per 1,000 contacts. In Excel: =SUM(F4:F13)/(SUM(B4:B13)-SUM(C4:C13)) for the positive-reply rate and =SUM(G4:G13)/SUM(B4:B13)*1000 for the meetings.
One bad list changes the whole quarter
The trade-fair list makes up almost a fifth of all contacts, and it drags every total down. Without it, the remaining nine campaigns have 8,250 contacts, a bounce rate of 3.4% (the old list alone: 9.5%) and 9.8 meetings per 1,000 contacts instead of 8.0. The cost of that one send is larger than the wasted effort. Hundreds of bounces in a few days are exactly the pattern that makes mail providers distrust a sending domain, and a distrusted domain delivers the good campaigns to the spam folder as well.
The opposite lesson comes from the smallest campaign. The referral look-alikes are only 500 contacts, yet they deliver 13 meetings, almost as many as the 1,200-contact 3PL campaign. A short, well-chosen list beats a long, careless one on every measure in the table. That is the thread running through the whole course: in cold outreach, quality of the list and relevance of the message decide the result, and volume mostly multiplies whatever quality is already there.
Common mistakes and how to spot them
- Dividing replies by opens.
=E4/D4produces an impressive number, but it rests on the unreliable open count and cannot be compared between campaigns. Use delivered emails as the base. - Dividing by contacts instead of delivered. For lists with few bounces the difference is small. For the trade-fair list the reply rate would be 1.35% instead of 1.49%, and the two problems, a bad list and a weak message, would be blended into one number.
- Missing brackets.
=D4/B4-C4divides first and then subtracts the bounces, which gives a large negative number. If an open rate is below zero, look for the brackets. - Calling every reply a success. A campaign that makes people angry enough to answer has a fine reply rate. Always report the positive-reply rate next to it.
- Counting emails instead of people. If a contact receives four emails and answers the third, that is one contact and one reply, not four sends and a 25% reply rate. Rates per step are a separate analysis (lesson 5).
- Averaging rates. Totals are built from summed counts. An average of rates is only right when every campaign has the same size.
- Reading too much into small numbers. Two meetings instead of one doubles the rate of a small campaign. Lesson 11 shows how to test whether a difference is larger than chance.
Rules and good manners come before arithmetic
No rate in this lesson says whether an email should have been sent in the first place. That question has to be answered before the campaign, not after it. The rules on unsolicited business email differ from country to country. Some allow a first email to a business contact as long as the sender is clearly identified, the content is not misleading and every email offers a simple way to opt out. Others require the recipient's prior consent, even between companies. Data-protection law adds duties of its own wherever personal data such as a named work address is processed.
Brackenmoor Labs therefore works with a few fixed principles, and the later lessons build them into the spreadsheets: contacts are chosen by their professional role and their company's business, never by private or protected characteristics; every list is checked against a suppression list of opt-outs, customers and earlier bounces; every email says who is writing and why; an opt-out is honoured quickly and for good; and each recipient country is checked for a consent requirement before anyone there is contacted.
Quick reference: the measures of this lesson
| Measure | Formula | Excel (row 4) | All campaigns |
|---|---|---|---|
| Delivery rate | (contacts − bounced) ÷ contacts | =(B4-C4)/B4 | 95.4% |
| Open rate | opened ÷ delivered | =D4/(B4-C4) | 42.6% |
| Reply rate | replied ÷ delivered | =E4/(B4-C4) | 4.41% |
| Positive-reply rate | positive replies ÷ delivered | =F4/(B4-C4) | 1.59% |
| Meetings per 1,000 contacts | meetings ÷ contacts × 1,000 | =G4/B4*1000 | 8.0 |
Where these numbers are used in practice
- Weekly campaign review. Export contacts, bounces, replies and meetings per campaign from the sending tool or the CRM, paste them into the sheet and compare the rates. Ten minutes are enough to see which list deserves the next wave and which should be stopped.
- Deciding where to invest. A low delivery rate calls for better data. A good delivery rate with a low reply rate calls for a better message or a better-matched audience. A good reply rate with few positive replies means the email provokes rather than interests.
- Planning. Meetings per 1,000 contacts turns a sales target into a list size. A team that needs 40 meetings and expects 8 per 1,000 contacts needs about 5,000 suitable contacts, which is often more than the market holds.
- Reporting to management. One line per campaign with delivery rate, positive-reply rate and meetings per 1,000 contacts says more than a page of activity counts.
What you will build in the rest of this course
Lesson 1 is free. Lessons 2 to 20 each have their own live lab with a new data set of Brackenmoor Labs and follow the workflow of an outreach programme from the list to the yearly dashboard. The first group completes the core of the topic.
- List Quality: Verification Results, Valid Share and the Expected Bounce Rate per Source. You compare twelve list sources by valid share, estimate the bounce rate of sending valid plus risky addresses and decide per source what may be sent.
- Suppression and List Hygiene: Duplicates, Opt-Outs, a Cooling-Off Period and the Sendable List. You flag duplicate addresses, count them, decide per contact between Suppress, Wait and Send, and count the sendable prospects and the unique addresses.
- Sending Volume Ramp-Up: Daily Limits per Mailbox, a Cap, Emails per Week and Cumulative Capacity. You build a ten-week ramp-up plan with a capped daily limit per mailbox, the weekly sending volume, the cumulative capacity and a chart.
- Sequence Steps: Reply Rate per Step, the Cumulative Reply Rate and the Value of Each Follow-Up. You measure an eight-step sequence step by step: reply rate per step, cumulative reply rate, share of positive replies and opt-outs per reply, with a chart.
- From Positive Reply to Meeting: Booking Rate, Show Rate and Meetings Held per 1,000 Contacts. You follow eight reps from positive replies to meetings booked and held, compare them per 1,000 contacts, rank them and calculate the team show rate.
- Cost per Meeting: Data, Labour and Tool Costs per Campaign and the Fully Loaded Cost per Meeting Held. You add data, labour and allocated tool costs for eight campaigns and calculate the cost per meeting held per campaign and in total.
- From Meetings to Pipeline: Opportunity Rate, First-Year Revenue and the Return on Outreach Cost. You follow the meetings of eight segments to opportunities and deals, calculate revenue, the return on outreach cost and the cost per deal, and chart the result.
The second group goes deeper into segments, deliverability, testing, planning and the question of when to stop.
- Segments with SUMIFS: Positive-Reply Rate and Meetings by Industry and by Job Title. You summarise sixteen campaign cells by industry and by job title with SUMIFS and build rates from sums, including a lookup with two criteria.
- Deliverability Checks: Bounce and Spam-Complaint Rates per Mailbox, Limits and an Action Flag. You calculate bounce and complaint rates for ten mailboxes, count the exceeded limits, set an action with IFS and highlight the problem mailboxes.
- Testing Two Email Versions: Difference in Reply Rate, Pooled Rate, z-Score and a Verdict. You evaluate eight A/B tests of cold emails with a z-test for two proportions and label each difference as real or noise.
- Reply Triage: Categories, Share of Replies, Working Days to Respond and Late Opt-Outs. You analyse a reply log: working days to handle each reply, replies and meetings per category, shares, and opt-outs processed later than the limit.
- Forecasting Meetings: Weekly Trend with SLOPE, FORECAST.LINEAR and the Volume Needed for a Target. You measure the weekly trend of booked meetings, forecast a future week, work back from a meeting target to the emails needed and chart the series.
- Capacity Planning: Contacts, Daily Emails, Mailboxes and Domains Needed for a Meeting Target. You work backwards from a monthly meeting target to contacts, emails per day, mailboxes, domains and their monthly cost, rounding up at every step.
- Personalisation Tiers: Research Cost, Positive-Reply Rate and Research Cost per Meeting by Tier. You cost three levels of personalisation, compare their positive-reply rates with SUMIFS and their research cost per meeting with SUMPRODUCT, and chart it.
- Send Timing: Reply Rates by Weekday and Time Block with SUMIFS, AVERAGEIFS and a Heat Map. You build a weekday-by-time grid of reply rates with one mixed-reference SUMIFS formula, add day totals and averages and shade the grid as a heat map.
- What-If Grid: Cost per Meeting for Combinations of Positive-Reply Rate and Cost per Contact. You build a two-way grid of the cost per meeting with mixed references, mark the combinations within the target and solve for the rate and cost limits.
- Data Sources Compared: Cost per Valid Contact, a Weighted Quality Score, Cost per Meeting and Rank. You compare eight data sources by cost per valid contact, a weighted quality score, cost per meeting and rank, and measure how concentrated the meetings are.
- List Fatigue: Wave-by-Wave Rates, an Index vs Wave 1, Negative Signals and a Stop Rule. You track eight waves to the same list, index the positive-reply rate to wave 1, compare negative signals and cost with positive replies and apply a stop rule.
- Final Project: Quarterly Outreach Dashboard with Positive-Reply Rate, Cost per Meeting, a Cap Flag and a Combo Chart. You roll twelve months up into a quarterly dashboard with meetings held, positive-reply rate and cost per meeting, flag the quarters above the cap and chart volume against cost.
Summary
A cold email campaign is a funnel from contacts emailed to meetings booked. The delivery rate, =(B4-C4)/B4, tells you whether the list is sound. Open, reply and positive-reply rates divide by the delivered emails, and of the three the positive-reply rate is the one to trust. Meetings per 1,000 contacts, =G4/B4*1000, sums the campaign up in a number that can be planned with. Totals come from summed counts. In the example, a small, well-chosen list produced 26.0 meetings per 1,000 contacts and an old, unverified one 0.5.
Now scroll up to the lab and build the five columns yourself. It takes about fifteen minutes, the Coach checks every step, and at the end you have a sheet you can fill with the counts of your own campaigns.
Frequently asked questions
How do I calculate the reply rate of a cold email campaign?
Divide the number of contacts who replied by the number of delivered emails, that is contacts emailed minus bounces. In Excel: =E4/(B4-C4) with replies in E, contacts in B and bounces in C. Count each person once, however many emails of the sequence they received.
What is the difference between reply rate and positive-reply rate?
The reply rate counts every answer, including refusals and requests to be removed. The positive-reply rate counts only answers that show interest or name the right colleague, divided by delivered emails. It is the better measure of how relevant the list and the message are.
Why should open rates in cold email be treated with caution?
Opens are counted when a small image in the email is loaded. Some mail programs load images automatically and others block them, so opens are both over- and under-counted. Use them as a rough warning signal and judge campaigns by replies and meetings.
Should rates be based on emails sent or emails delivered?
Open, reply and positive-reply rates are based on delivered emails, because a bounced email could not be answered. The delivery rate and the meetings per 1,000 contacts are based on all contacts emailed, because every contact costs money and uses up a name on the list.
How do I calculate meetings booked per 1,000 contacts?
Divide the meetings booked by the contacts emailed and multiply by 1,000, in Excel =G4/B4*1000. The number makes campaigns of different sizes comparable and turns a meeting target into the list size that is needed.
Is cold emailing business contacts allowed?
That depends on the country of the recipient. Some countries allow a first business email with a clear sender identity, honest content and an easy opt-out; others require prior consent even between companies, and data-protection rules apply as well. Check the rules for each country and honour every opt-out. This course is not legal advice.
Keep going: Cold Email Outreach for B2B Prospecting: List Quality, Sending Ramp-Up, Reply Rates, Sequences, Meetings Booked & Cost per Meeting
This free lab is lesson 1 of 20. Unlock the other 19 hands-on labs, the final exam and your certificate for a one-time $9.90.
- List Quality: Verification Results, Valid Share and the Expected Bounce Rate per Source
- Suppression and List Hygiene: Duplicates, Opt-Outs, a Cooling-Off Period and the Sendable List
- Sending Volume Ramp-Up: Daily Limits per Mailbox, a Cap, Emails per Week and Cumulative Capacity
- Sequence Steps: Reply Rate per Step, the Cumulative Reply Rate and the Value of Each Follow-Up
- From Positive Reply to Meeting: Booking Rate, Show Rate and Meetings Held per 1,000 Contacts
- Cost per Meeting: Data, Labour and Tool Costs per Campaign and the Fully Loaded Cost per Meeting Held
- From Meetings to Pipeline: Opportunity Rate, First-Year Revenue and the Return on Outreach Cost
- Segments with SUMIFS: Positive-Reply Rate and Meetings by Industry and by Job Title
- Deliverability Checks: Bounce and Spam-Complaint Rates per Mailbox, Limits and an Action Flag
- Testing Two Email Versions: Difference in Reply Rate, Pooled Rate, z-Score and a Verdict
- Reply Triage: Categories, Share of Replies, Working Days to Respond and Late Opt-Outs
- Forecasting Meetings: Weekly Trend with SLOPE, FORECAST.LINEAR and the Volume Needed for a Target
- Capacity Planning: Contacts, Daily Emails, Mailboxes and Domains Needed for a Meeting Target
- Personalisation Tiers: Research Cost, Positive-Reply Rate and Research Cost per Meeting by Tier
- Send Timing: Reply Rates by Weekday and Time Block with SUMIFS, AVERAGEIFS and a Heat Map
- What-If Grid: Cost per Meeting for Combinations of Positive-Reply Rate and Cost per Contact
- Data Sources Compared: Cost per Valid Contact, a Weighted Quality Score, Cost per Meeting and Rank
- List Fatigue: Wave-by-Wave Rates, an Index vs Wave 1, Negative Signals and a Stop Rule
- Final Project: Quarterly Outreach Dashboard with Positive-Reply Rate, Cost per Meeting, a Cap Flag and a Combo Chart