Link positions: where the links sit and how dense they are

What this screen is

The app calls this screen Link position & Density, under the sub-line Distribution of links within newsletters - position and density analysis. Two questions sit behind it. How heavily linked is a newsletter relative to the amount of copy it carries, and how far into the email does a reader get before the first clickable element appears.

Both numbers come from counting. Every link with a destination in the email HTML counts, and so does every word of body text. Link density is those two numbers combined into links per 1000 words. First link position is the character offset of the first link expressed as a percentage of the whole email HTML. No judgement about link strategy goes into either one, with a single exception noted below.

Under the middle tile and above the histogram sits a best-practice band of 7 to 15 links per email, with 10 as the ideal. That band is fixed editorial guidance. Nobody measured your peers to get it, and it never moves when your data moves.

The chart at the bottom pairs raw link count against an estimated click-through rate. Newsletrix receives competitor newsletters as an ordinary subscriber, so it can never see another sender's real clicks. The rate on that axis is an estimate made from the content of each email, so any relationship the chart shows runs between something counted and something estimated. That card also ignores the Brand and date filters on this screen and always plots up to your 500 most recent newsletters.

One thing the screen never does is list the individual links. You cannot see from here which URLs, which sections or which images produced a density figure. No per-newsletter list, no export, no sort control.

Open it when

• You are rewriting a template and want to know whether your emails carry too many or too few links against the 7 to 15 band.
• A competitor's newsletter feels linky and you want to quantify it: filter to that brand and read the average density and the histogram.
• You suspect your first call to action sits too far down the email and want to see how many sends put the first link in the top quarter versus the bottom quarter.
• You are testing a fewer links, one CTA hypothesis and want to see whether link-heavy emails show any relationship with estimated CTR.
• You want a link profile for one specific window, such as a holiday campaign period, rather than your whole archive.
• You are briefing a designer and need a number for how far into the email the first clickable element appears.

Screen anatomy

The map below runs top to bottom, with sample numbers. Markers match the list underneath.

1. Average link density The headline number of the screen: mean links per 1000 words across every newsletter that passes the filters. The hint under it reads links per 1000 words.
2. Average link count Mean links per email, with a benchmark strip underneath showing your value against the fixed 7 to 15 best-practice band and a chip reading in range, above range or below range. This is the only tile the band applies to.
3. Analyzed emails The sample size behind the two tiles to its left. Single-digit values mean you are reading anecdote.
4. First link position distribution A doughnut with exactly four slices, Q1 (0-25%) through Q4 (75-100%), counting how many newsletters place their first link in each quarter of the email HTML.
5. Link density (histogram) One bar per density band: 0-10, 10-20, 20-30, 30-40 and 40+ links per 1000 words. The grey strip above the bars is the 7 to 15 links per email band, which is a different unit from the bars.
6. Link count vs. Est. CTR One dot per newsletter, a fitted trend line, an amber dashed vertical line at the ideal of 10 links, and a grey dashed horizontal line at the industry average CTR. The badge in the title row carries the correlation value, a strength word and the number of points.
7. Section header The title and its sub-line, plus a round help icon on the right that opens a read-only explainer called Strategic Link Placement. It changes no data.
8. Filter card Brand, From date and To date, together in one bordered card. No Apply button; changing any control redraws the tiles and both charts immediately.
9. Loading and empty states While data loads, everything below the filters is replaced by a single spinner rather than greyed-out cards. A chart with nothing to plot prints No data available, and the scatter prints its own message with the number of points it found.

Every panel explained

Average link density

The mean number of links per 1000 words of body copy across everything the current filter covers.

How it is calculated. Each newsletter gets its own ratio first: link count multiplied by 1000, divided by body word count. Those per-email ratios are then averaged and rounded to two decimals. A newsletter only qualifies if it has a known link count and at least one word of body text. An email with no links at all drops out of the average rather than counting as zero.

How to read it. A 1500-word newsletter with 15 links scores 10. High numbers mean link-dense, list-style content such as a round-up or a digest. Low numbers point to long-form editorial with few places to click.

Average link count

The mean number of links per email, with the best-practice band printed underneath as a strip reading your value, then Best practice 7-15, then a coloured chip.

How it is calculated. The link counts of all qualifying newsletters are summed and divided by how many there are, rounded to two decimals. Emails with no links are excluded. A link is any element in the email HTML that carries a destination, which means the logo, the view in browser link, every social icon, each footer legal link, the unsubscribe link and any image wrapped in a link all count. The strip appears whenever the average is above zero and always ends in a chip. Green and reading in range when your average sits inside 7 to 15. Red and reading above range or below range when it falls outside.

How to read it. Compare the number to the 7 to 15 band with 10 as the ideal. The reasoning behind the band, as the app states it, is that below 7 links you under-use the email real estate, and above 15 you dilute click intent and raise spam-filter flags.

Analyzed emails

How many newsletters made it into the density calculation under the current filter.

How it is calculated. A straight count of the newsletters behind the two tiles to its left. Each one needs a known link count, at least one word of body text, a finished analysis, and a match on the Brand, date and workspace filters.

How to read it. Treat it as the confidence dial for the whole top row. Single-digit values make both averages unstable; a double-digit sample per brand is the practical minimum before you quote a density figure to anyone.

First link position distribution

A doughnut counting how many newsletters put their first link in each quarter of the email, labelled Q1 (0-25%), Q2 (25-50%), Q3 (50-75%) and Q4 (75-100%).

How it is calculated. First link position is the character offset of the first link divided by the total length of the email HTML, times 100, to two decimals. The screen sorts that percentage into quarters. 25 or below goes to Q1, above 25 up to 50 to Q2, above 50 up to 75 to Q3, and anything above 75 to Q4. A newsletter reaches the doughnut only if it has at least one link and some body text, and a position of exactly 0.00 lands in no slice.

How to read it. A fat Q1 slice means something clickable appears near the top of every send, which in most templates is the logo, the view in browser link or a hero button. A fat Q3 or Q4 slice means readers scroll past most of the email before anything responds to a tap, which is worth checking against the mobile preview of that template.

Link density (histogram)

How many newsletters fall into each links-per-1000-words band, across five buckets: 0-10, 10-20, 20-30, 30-40 and 40+. The series is labelled Number of Emails.

How it is calculated. Each qualifying newsletter's density is placed in a bucket with edges at 10, 20, 30 and 40. The upper edge is open, so a density of exactly 10 lands in the 10-20 bucket, not in 0-10. Buckets with no newsletters are left out entirely rather than drawn as a flat zero bar.

How to read it. The histogram is the spread that the average tile hides. A distribution clustered in 0-10 means copy-heavy emails; a long tail into 40+ means short link-list formats such as round-ups and digests. Two brands can share an average of 18 while one is tightly clustered and the other swings between 5 and 45.

Link count vs. Est. CTR

A full-width chart with one dot per newsletter, link count on the horizontal axis and estimated CTR on the vertical axis, described in the app as: Do more links lift CTR, or dilute it? The orange dashed line marks the best-practice ideal of ~10 links; the grey line is industry-average CTR.

How it is calculated. The chart plots your 500 most recent newsletters that have both a link count and an estimated CTR, with an ordinary least squares trend line across them and a Pearson correlation in the badge. No correlation is shown when there are fewer than three points or when one of the axes has no variation. The badge turns the value into a word. 0.5 or above is strong, 0.3 or above is moderate, 0.1 or above is weak, and below 0.1 reads as no correlation. The amber dashed vertical line sits at 10 links, the ideal from the best-practice band. The grey dashed horizontal line is the median CTR for the benchmark industry selected in your account, which defaults to retail at 2.01 percent.

How to read it. Look at the badge before the dots. A downward trend that crosses the amber line suggests that, in your tracked set, links beyond roughly ten are not associated with a higher estimated CTR. A flat or noisy cloud is the most common and most honest outcome, and it means link count is not the variable driving anything here.

When the screen has nothing to show

Each panel has its own qualifying rules, so parts of the screen can be empty while others are full. The example below shows a new account with a handful of collected newsletters.

How it is calculated. A newsletter reaches this screen only after Newsletrix has finished analysing it. The tiles and the histogram then need a known link count and at least one word of body text, the doughnut needs at least one link and a first link position above zero, and the scatter needs at least three newsletters that have both a link count and an estimated CTR, with variation on both axes.

How to read it. Plain-text-only emails end up with a link count of zero, so they add nothing to the averages and appear in no slice. A screen that is full at the top and empty at the bottom means you have enough collected newsletters but not yet enough with an estimated CTR attached.

Metric definitions

Filters and controls

A few controls you may go hunting for are deliberately absent. There is no reset button, no export of any kind and no sort or search control. Nor is there an Apply button, because every filter change redraws the screen at once. And the scatter at the bottom ignores Brand, From date and To date entirely.

How to use it

1. Check the sample first Read Analyzed emails before anything else. Under about ten newsletters, both averages are anecdote. Widen the date range until the count is comfortably in double digits.
2. Compare Average link count to the band This is the only number the 7 to 15 band belongs to. If the chip says above range, you are carrying more destinations per email than the guidance suggests, boilerplate links included.
3. Read the histogram for spread, not level The bars tell you whether your set is consistent or bimodal. A brand that splits between a 0-10 bar and a 30-40 bar is running two different email formats, and averaging them describes neither.
4. Look at the quartiles for the first click If Q3 and Q4 together hold a meaningful share of your sends, readers are scrolling through most of the email before anything is clickable. Check whether that is a deliberate long read or an accident of the template.
5. Filter to one competitor and repeat Select a brand, keep the date range, and compare the same four readings against your own. Note the numbers somewhere, because the screen keeps no history.
6. Use the scatter only as a hypothesis generator Read the badge, remember that the vertical axis is an estimate and that the chart ignores your filters, then test the idea on your own list where you can see real clicks.

Limits and caveats

• Link count includes every link with a destination in the email HTML, which means the logo, view in browser, social icons, footer legal links, unsubscribe and linked images. It is neither a count of distinct destinations nor a count of calls to action.
• First link position is measured in raw HTML characters, so it is a rough proxy for scroll depth. Markup-heavy templates report a different percentage than lean ones that look identical.
• An email with no links disappears three times over: from the density average, from the link count average and from the doughnut. It never shows up on this screen as a zero.
• The doughnut and the tiles use different qualifying rules, so the four slices need not add up to Analyzed emails.
• Both averages are unweighted means, so a 300-word teaser counts exactly as much as a 3000-word edition.
• The 7 to 15 band and the ideal of 10 are prescriptive guidance dated January 2024, not a measurement of your peers, and they never change with your data.
• The scatter plots an AI-estimated CTR, is capped at the 500 most recent qualifying newsletters, fits a simple line with no outlier handling and no significance test, and ignores the Brand and date filters.

FAQ

Related screens