Why link building needs a funnel
Ask five marketers how their link building is going and you will often hear a single number: "we won 20 links last month". It sounds like a result, but it explains nothing. Were those 20 links the harvest of 150 carefully chosen emails or of 3,000 copy-and-paste messages? Did most prospects never open the email, or did they open it, read the pitch and decide it was not for them? Without the steps in between you cannot tell a good month from a lucky one, and you cannot tell what to fix.
Link building is the work of earning links from other websites, and most of it is done by outreach: writing to an editor, a blogger, a journalist or the keeper of a resource page and offering something their readers would benefit from. Outreach is a process with clear stages, and a process can be measured. The measuring tool is the outreach funnel: emails sent, emails opened, replies received, links won. This free lesson builds that funnel in a spreadsheet, turns it into three rates and shows how to read them.
The example is Quillfern Outdoors, a fictional online shop for hiking and camping gear. Its outreach team ran ten campaigns in the first half of 2026. All companies, sites, people and numbers are made up for this course, and any benchmark you read here is illustrative, not an industry standard. The table below is the same data as in the interactive lab above, so you can follow every figure cell by cell.
The four stages: sent, opened, replied, linked
Every outreach email travels through the same four stages, and at each stage some emails drop out.
- Sent. The number of prospects who were emailed in the campaign. This is the base of the funnel and the denominator of every rate in this course.
- Opened. Emails that the recipient opened, as far as the outreach tool can tell. Tools detect opens with a tiny tracking image, which some mail programs block and others load automatically, so the count is an estimate in both directions.
- Replied. Prospects who wrote back. A reply can be a yes, a question, a price list or a polite no; at this stage all of them count, because a reply means a human being read the pitch.
- Linked. Prospects whose site now carries a link to yours. This is the result the campaign exists for, and the only stage you can verify yourself by looking at the page.
Three rates with one denominator
Counts alone cannot be compared between a campaign of 60 emails and one of 420. Rates can. In this course all three funnel rates are divided by the same number, the emails sent:
- Open rate = opened ÷ sent
- Reply rate = replies ÷ sent
- Link rate = links won ÷ sent
You will meet other definitions in tools and articles: reply rate per opened email, or links per reply. They are legitimate step-to-step rates and useful for diagnosis, but they must not be mixed with the per-sent rates in one table. A reply rate of 20 % per open and a reply rate of 10 % per email sent can describe the very same campaign. Decide on one definition, write it into the column header, as the lab does with Link rate (links ÷ sent), and keep it for the whole report.
A fully worked example: ten campaigns at Quillfern Outdoors
Here are the ten campaigns with the four counted columns and the three calculated rates, exactly as they appear when the lab is solved.
| Campaign | Sent | Opened | Replies | Links | Open rate | Reply rate | Link rate |
|---|---|---|---|---|---|---|---|
| Trail guide resource pages | 180 | 97 | 22 | 9 | 53.9 % | 12.2 % | 5.0 % |
| Gear review bloggers | 240 | 118 | 31 | 8 | 49.2 % | 12.9 % | 3.3 % |
| Hiking habits survey (digital PR) | 420 | 189 | 38 | 21 | 45.0 % | 9.0 % | 5.0 % |
| Broken-link replacements | 150 | 84 | 19 | 11 | 56.0 % | 12.7 % | 7.3 % |
| Podcast guest pitches | 90 | 61 | 14 | 5 | 67.8 % | 15.6 % | 5.6 % |
| Local hiking clubs | 130 | 88 | 27 | 12 | 67.7 % | 20.8 % | 9.2 % |
| Unlinked brand mentions | 60 | 44 | 23 | 16 | 73.3 % | 38.3 % | 26.7 % |
| Campsite map data study | 380 | 160 | 29 | 14 | 42.1 % | 7.6 % | 3.7 % |
| Expert quotes for journalists | 260 | 109 | 17 | 6 | 41.9 % | 6.5 % | 2.3 % |
| Supplier and partner pages | 70 | 52 | 20 | 13 | 74.3 % | 28.6 % | 18.6 % |
| Total | 1,980 | 1,002 | 240 | 115 | 50.6 % | 12.1 % | 5.8 % |
The formulas are short. They are written for row 4, the first campaign, and copied down to row 13:
- Open rate (F4):
=C4/B4, opened divided by sent. - Reply rate (G4):
=D4/B4, replies divided by sent. - Link rate (H4):
=E4/B4, links won divided by sent. - Totals (B14:E14):
=SUM(B4:B13)in B14, copied to the right as far as E14. - Overall link rate (H14):
=E14/B14, or straight from the data=SUM(E4:E13)/SUM(B4:B13).
Take the second campaign, the gear review bloggers. 240 emails were sent and 118 were opened, so the open rate is 118 ÷ 240 = 49.2 %. 31 bloggers replied, a reply rate of 12.9 %, and 8 links were won, a link rate of 3.3 %. Now the broken-link campaign: 150 emails, 19 replies, reply rate 12.7 %, and 11 links, link rate 7.3 %. Telling a site owner that a link on their page leads nowhere, and offering a working replacement, is a small favour, and the numbers show it.
Step by step: the totals and the overall link rate
Add up each counted column with SUM. The programme sent 1,980 emails, 1,002 were opened, 240 were answered and 115 links were won. Dividing each total by the 1,980 emails gives the programme funnel: an open rate of 50.6 %, a reply rate of 12.1 % and an overall link rate of 5.8 %. In plain words: on average the team had to send about 17 emails for one link.
It is tempting to get the overall figure by averaging the ten campaign link rates. Do not. The plain average of the ten rates is 8.7 %, far above the true 5.8 %. The reason is weight. The unlinked-mentions campaign has a link rate of 26.7 % but only 60 emails; the expert-quote campaign has 2.3 % on 260 emails. A plain average counts both equally, although one is more than four times the size of the other. Dividing the total links by the total emails weights every campaign by its size automatically.
Reading the chart: which campaigns work?
A look at the link rates campaign by campaign shows a pattern that repeats in almost every outreach programme.
The highest link rates belong to the warm lists. Unlinked brand mentions (26.7 %) are writers who already named Quillfern and only need to add the link. Supplier and partner pages (18.6 %) are businesses with an existing relationship. Both lists are small, 60 and 70 emails, and they cannot be made much larger: there are only so many mentions and partners. The lowest rates belong to the cold pitches to journalists and bloggers, who receive many such emails every day.
Yet the rate is not the whole story. The hiking habits survey has a modest link rate of 5.0 %, below the programme average, and still won 21 links, more than any other campaign, because its list was large. Later lessons add two more views that change the ranking again: what each link cost (lesson 3) and how good the linking sites are (lesson 7). A link from a national news site and a link from a partner's footer are both one link in this table.
What each rate tells you to fix
The value of the funnel is that each rate points to a different part of the work. Compare a campaign's three rates with the programme row and ask where it falls behind.
- Low open rate: the email did not arrive or did not look worth opening. Check the sending domain's reputation, the bounce rate of the list, the subject line and the sender name. Remember that the open count itself is unreliable before you draw strong conclusions.
- Good opens, low reply rate: people read the pitch and it did not concern them. The usual causes are a prospect list that does not fit the topic and an email that talks about the sender's wishes instead of the reader's audience.
- Good replies, low link rate: the conversation started and then stalled. Perhaps the linked page is thin, the site asks for payment, or nobody answered the editor's question in time. A reply that sits unanswered for a week is a lost link.
- Everything low: usually the list. No template rescues a campaign aimed at sites that have no reason to care.
In the example, the expert-quote campaign has an open rate of 41.9 %, somewhat below the programme's 50.6 %, but a reply rate of only 6.5 %. The emails were seen and ignored. That points to the pitch and the choice of journalists, not to deliverability. For the whole programme, 240 replies led to 115 links, so almost every second reply ended in a link (47.9 % of replies). The bottleneck of this programme is clearly getting an answer, not converting it.
Common mistakes with outreach metrics
The arithmetic is simple, which makes the mistakes easy to overlook.
- Mixing denominators. Open rate per sent, reply rate per open, link rate per reply: each number looks impressive and none can be compared. Use emails sent for all three, and label step-to-step rates clearly when you add them.
- Averaging rates. The plain average of campaign rates overstates the programme when small warm lists perform well. Divide totals by totals.
- Counting emails instead of prospects. A prospect who received a first email and two follow-ups is one prospect. If the denominator counts three emails, every rate shrinks to a third and campaigns with more follow-ups look worse than they are.
- Trusting opens. Deciding between two subject lines on a difference of a few opens is deciding on noise. Lesson 13 shows how to test whether a difference in replies is real.
- Reading small numbers as facts. With 60 emails, one link more or less moves the link rate by 1.7 points. A rate from a small campaign is a rough estimate.
- Forgetting quality and cost. A high link rate from weak, irrelevant sites is not a success, and a rate says nothing about the hours behind it.
Outreach with good manners
A funnel measures people's reactions to email they did not ask for, so the way you send matters as much as the numbers. Write only to people for whom the topic is part of their published work, and show in the first two lines that you have read it. Say who you are and what you are asking. Offer something of use to their readers, not a favour to you. Do not pay for links unless the site labels them as sponsored, since buying links to manipulate rankings is against search-engine guidelines. Follow up once or twice, then stop. And when someone asks not to be contacted again, record it and honour it in every later campaign.
These habits are not a cost on top of the funnel, they are what moves it. Relevance raises the reply rate, an honest and useful page raises the link rate, and a sender whom editors recognise and do not resent keeps a healthy open rate over the years. Rules for unsolicited business email also differ from country to country, so a team should know which ones apply to it.
How to use the funnel week by week
A funnel is most useful as a habit. Keep one row per campaign, or per weekly batch, and fill in the four counts when the batch is two to three weeks old, because replies and links arrive with a delay. Compare each new row with the programme row, not with the best campaign ever. When a rate falls behind, change one thing in the next batch, the list, the subject line or the pitch, and watch the matching rate. Changing everything at once leaves you with a new number and no idea what caused it.
Set expectations with your own history, not with numbers from the internet. Link rates depend on the industry, the brand, the kind of prospect and the quality of the page you offer, and published "average reply rates" are drawn from senders who may have little in common with you. Your own programme row, tracked over a few quarters, is the benchmark that counts.
What the rest of the course covers
The funnel is the first of twenty labs, all with Quillfern Outdoors and all checked cell by cell as you type.
- Lessons 2 to 8, the core: qualifying a prospect list with AND and COUNTIFS, cost per link including the team's hours, the anchor-text mix against cautious guide values, referring-domain growth with a running total, the gap to competitors and the months needed to close it, a weighted link quality score with tiers, and follow-up sequences.
- Lessons 9 to 14, going deeper: results by prospect type with SUMIFS, digital PR coverage with unlinked mentions and link reclamation, a trend forecast with SLOPE and FORECAST.LINEAR, link decay and loss rates, a significance test for two outreach templates, and list hygiene with opt-outs and a cool-down period.
- Lessons 15 to 19, decisions: a budget ranked by cost per quality point, a what-if grid with mixed references, the payback and ROI of link campaigns, a risk audit of inherited links and a capacity plan that works backwards from target links to hours.
- Lesson 20, the final project: a half-year dashboard by tactic with a combo chart, followed by a 20-question exam and a certificate.
You need no special software and no SEO tool subscription: every lab runs in the browser with Excel-style formulas, and each formula you learn works the same way in Excel or Google Sheets. The data are fictional throughout, so that you can concentrate on the logic and carry it over to your own campaign log afterwards.
Frequently asked questions
What is the outreach funnel in link building?
It is the sequence every outreach email passes through: emails sent, emails opened, replies received and links won. Dividing each stage by the emails sent gives the open rate, reply rate and link rate, which show where a campaign loses prospects.
How do I calculate the link rate of an outreach campaign in Excel?
Divide the links won by the emails sent, for example =E4/B4, and format the cell as a percentage. For several campaigns together use =SUM(E4:E13)/SUM(B4:B13) instead of averaging the campaign rates.
Why should I not average the link rates of my campaigns?
A plain average gives a campaign of 60 emails the same weight as one of 420. In the worked example the average of the ten campaign rates is 8.7 %, while the true overall link rate, total links divided by total emails, is 5.8 %.
Are email open rates reliable for outreach?
Only roughly. Opens are detected with a tracking image that some mail programs block and others load automatically, so the count can be too low or too high. Use the open rate as a rough deliverability signal and base decisions on replies and links.
What is a good reply rate or link rate for outreach?
There is no universal figure. Rates depend on the industry, the brand, the kind of prospect and the page you offer. Warm lists such as unlinked mentions convert far better than cold pitches. Your own programme average over several quarters is the most useful benchmark.
Should the reply rate be divided by emails sent or by emails opened?
Both are used. This course divides every funnel rate by emails sent, so the rates can be compared with one another. Replies per opened email is a step-to-step rate; it is useful for diagnosis but should be labelled clearly and not mixed with per-sent rates.
Keep going: Link Building & Digital PR Outreach: Prospect Lists, the Outreach Funnel, Cost per Link, Anchor-Text Mix, Link Quality & Referring-Domain Growth
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.
- Building a Prospect List: Qualifying Sites by Authority, Organic Traffic and Topical Relevance with AND and COUNTIFS
- Cost per Link: Labour, Content and Tool Costs per Campaign, the Blended Cost per Link and a Bar Chart
- Anchor-Text Mix: Branded, Naked, Generic, Partial and Exact Match, Shares Against Guide Values and a Pie Chart
- Referring-Domain Growth: New and Lost Domains, Net Change, a Running Total and the Growth Rate of the Year
- Referring Domains vs Competitors: Growth %, Monthly Velocity, Rank and the Months Needed to Close the Gap
- Link Quality Scoring: Weighted Authority, Relevance, Traffic and Follow Status, Tiers with IFS and a Highlight Rule
- Follow-Up Sequences: Replies per Step, Reply Rate per Prospect, the Follow-Up Share and a Stacked Column Chart
- Results by Prospect Type: SUMIFS Summaries for Journalists, Bloggers, Resource Pages and Podcasts, and the Best Batch
- Digital PR Coverage: Follow Links, Nofollow Links and Unlinked Mentions, the Linked Coverage Rate and Link Reclamation
- Forecasting Referring Domains: SLOPE as Link Velocity, FORECAST.LINEAR, the Month a Target Is Reached and R²
- Link Decay: Link Age in Days, Loss Rates by Source Type, Surviving Old Links and a Rule That Highlights Lost Links
- Testing Two Outreach Templates: Lift in Reply Rate, the Pooled z-Test, p Values and a Verdict for Small Lists
- List Hygiene Before a Send: Duplicate Sites, Email Verification, Opt-Outs and a Cool-Down Period with Dates
- Allocating the Link Budget: Cost per Quality Point, Ranking Tactics and Funding in Rank Order Until the Budget Ends
- What-If Grid: Cost per Link by Emails Sent and Link Rate with Mixed References, a Heat Map and Backward Planning
- Do the Links Pay? Incremental Organic Visits, Margin per Month, Payback and 12-Month ROI per Target Page
- Link Risk Audit: Risk Points for Weak Sites, Exact-Match Anchors and Site-Wide Links, Labels with IFS and a Flag Rule
- Backward Planning: From Target Links to Replies, Emails and Hours with ROUNDUP, and the Capacity the Team Needs
- Final Project: A Half-Year Link Building Dashboard with Links, Link Rate, Cost per Link and Link-Weighted Quality by Tactic