First, define the engagement measure for each channel
"Engagement" means different things on the website, in the inbox and in a social feed, and each tool defines it in its own way. On the site, analytics tools count an engaged session by their own documented rules; the Google Analytics reports and dimensions guide sets out how Engagement RateShare of sessions that were engaged; on social, interactions relative to reach. is calculated there. In email, clicks and replies matter more than opens. On social platforms, saves, shares, comments and click-throughs each say something different. Choose one or two measures per channel, write their definitions down, and check whether any definition changed during the period you are judging; a redefined metric can look exactly like a collapse in interest.
Then match the measure to the job each piece was given. Kaushik's See, Think, Do, Care model is a useful guide: a See piece aimed at people with no current need succeeds when it is noticed and remembered, a Think piece when people read deeply and come back. Judging a Think guide on shares, or a See post on reading depth, produces a false alarm. If the measures are sound and the content still earns little attention, work through the table below; if the problem is visits without outcomes rather than attention, the content marketing troubleshooting hub routes you to the conversion page.
Diagnosis table
Diagnosis
| Symptom | Check | Likely cause | Fix |
|---|---|---|---|
| Pieces describe products, launches and company news; readers skim and leave | What share of recent pieces start from a reader's question | Written about the company, not the reader | Start every piece from a researched audience question |
| A burst of activity on publication day, then nothing | Where each piece appeared after its first announcement | Posted once, never distributed | Plan a sequence of appearances over weeks |
| Link posts and long captions get little reaction on a feed built for video or images | The native format of each channel against what was posted | Wrong format for the channel | Re-cut each piece into the channel's format |
| Readers arrive and leave within seconds; few reach the second section | Opening, headings and paragraph structure on a phone | Hard to scan on screen | Front-load the point; rewrite headings and paragraphs |
| Engagement spikes after a flurry of posts, then fades through silent weeks | Posting dates over recent months against the planned rhythm | Irregular cadence | Set a rhythm the team can keep in its busiest month |
Fix, verify and prevent each cause
Content about the company instead of the reader
Announcements, feature lists and company milestones interest the company. Readers engage with pieces that help them with something they are trying to do, in the words they use for it.
Fix: rebuild the topic list from the question bank described in the audience research guide, drawn from customer interviews, sales calls, support logs and the communities the audience uses, and rewrite headlines in the reader's phrasing rather than the brand's.
Verify: pieces built from researched questions earn more reading time, saves and replies than the company-led pieces they replace over the next few months.
Prevent: require every brief to quote the reader question it answers and its source, and keep a standing intake of new questions from sales and support.
Announced once and then forgotten
A piece shared once reaches the small part of the audience that happened to be looking that day. Most of the people it was written for never see it.
Fix: plan each piece's distribution before publication, as the content distribution channels guide describes: internal links from pages already read, the relevant newsletter segment, adapted posts released over several weeks, and genuine answers in the communities where the question is asked. The guide to sharing content in online communities covers reading each community's rules, answering in the thread first and saying who you work for.
Verify: each piece shows engagement from several sources over its first weeks rather than one spike on publication day.
Prevent: write the distribution notes and their owners into the brief, and do not publish a piece that has none.
The wrong format for the channel
Every platform is built around a native format and rewards it; a link with a headline on a video-first feed, or a long article pasted into a caption, looks imported and is treated that way by the audience and the ranking. The social content formats guide explains what each format is for.
Fix: re-cut each piece into the format of the channel it will appear on, such as a short clip, a carousel or a single text post making one point, and invite readers' own examples and responses, a kind of UGCContent about a brand made by customers or creators rather than by the brand itself., where the format allows.
Verify: saves, shares and comments per post rise for the adapted formats compared with the link posts they replaced.
Prevent: decide the format per channel at the brief stage and keep a short list of the formats each channel has rewarded.
Hard to read on a screen
Nielsen Norman Group's research on how users read on the web found that most people scan rather than read word by word, attending to headings, the start of paragraphs and anything that stands out. A piece that opens with background and hides its point in long paragraphs loses them in the first screen.
Fix: edit with the pass-by-pass routine in the writing for the web guide: answer first, descriptive headings, one idea per paragraph with the point in its first sentence, lists where items are parallel. For newsletters and reports, ordering the whole piece as an inverted pyramid applies the same answer-first rule.
Verify: the share of readers who scroll past the first section and the reading time on long pieces rise after the edit.
Prevent: keep the headings pass and the first-sentence pass in every edit, and read every draft on a phone before approval.
No rhythm the audience can learn
An audience forms a habit around a brand that appears when expected. Bursts of posting followed by silence teach the opposite, and feeds learn the inconsistency too.
Fix: set the cadence from the hours the team really has, subtracting replies and review first, as the brand voice and cadence guide sets out, and batch production so the rhythm holds through busy weeks.
Verify: engagement per post holds steady or rises across a full quarter of regular posting, and returning readers and repeat engagers increase.
Prevent: review the cadence against its KPIA metric tied directly to an objective. each quarter, raising it only when engagement per piece holds and lowering it when it falls.
A prevention routine for engaging content
Engagement problems are cheapest to prevent at the brief, so the routine starts there. Every brief names the reader question and its source, the channels the piece will appear on and the format for each, the sequence of appearances after publication, and the engagement measure each channel will be judged on. Before approval, a reviewer reads the draft on a phone, headings and first sentences only, and sends it back if the point is not clear from those alone. Weekly, the team checks that the planned rhythm was kept on each channel and that distribution steps scheduled for older pieces went out. Monthly, engagement is read by channel and by stage, per piece rather than in total, so that a weak format or a weak topic shows up instead of being averaged away. The pieces that earned the most attention are examined for what they had in common, and the next month's briefs borrow it. Quarterly, refresh the question bank from sales, support and community threads, retire formats a channel has stopped rewarding, and review the cadence against what the team actually produced.
A garden centre's blog and social posts wake up
A garden centre published 3 blog posts a month, mostly about new stock and events, shared each as a link on two social accounts on the day it went live, and posted nothing else in between. Posts averaged 4 reactions, and blog posts averaged 35 seconds of engaged time. A review of 200 customer questions from the till and the enquiry inbox produced a list of seasonal problems such as slugs, frost damage and shady borders. The centre switched to 2 problem-led guides a month, each opening with the answer, cut into a short video and a 5-slide carousel released over 3 weeks, posted on fixed days, and added to the monthly newsletter. Within a quarter, posts averaged 31 reactions and 6 shares, guides averaged 2 minutes 40 seconds of engaged time, and customers began sending photos of their own borders.
Common mistakes
Publishing company news and calling it content. Sharing each piece once on the day it goes live. Posting the same link and caption to every channel. Opening articles with background. Posting in bursts when there is time. Judging every piece on the same engagement number regardless of its job.
