When paying to promote a piece is worth it
The owned, earned and paid media guide names the property of paid media that matters most here: it is fast and scalable, and it stops when the money stops. That is why the content distribution channels guide puts paid promotion after search, email, the organisation's own pages and the social and community routes, and gives it one section and a few rows of its channel table. This tutorial turns that section into a sequence for a single piece, from choosing it to deciding whether to keep paying.
The channels guide names three situations in which paying makes sense, and they are the first test for any piece:
- Acceleration. Search, email or organic social already shows that readers find the piece, read it and act on it, and promotion takes it to a larger group of the same kind of people.
- Capture. The page answers a question people search for but does not yet rank for it, so paid search on that query brings searchers to it while the organic work catches up.
- A defined audience. The piece was written for a group no owned or earned channel reaches, and a paid placement is the only way in front of them.
Behind all three sits the guide's warning: paying multiplies a piece's existing pull and seldom rescues one that failed organically. If a distributed piece still earns few readers, start with the diagnosis for content that gets no engagement; if it draws readers who never take a next step, start with blog traffic that is not converting, and promote once the fix shows in the piece's own figures.
Two objective pages put limits on the money. The guide to content for consideration keeps promotion money for pieces whose readers demonstrably return or that appear in journeys ending in a conversion. The guide to content for lead generation judges paid promotion of gated assets on the qualified leads each produces, and whether an asset belongs behind a form at all is a decision the gated versus ungated content comparison works through.
Other pages own the neighbouring routes. A campaign to make the brand known among people not looking for the category is planned on reach and frequency, as the guide to paid social for brand awareness describes. Space in someone else's publication is a sponsored article or a newsletter sponsorship. A post written by an employee or a creator follows the permissions in the thought leader ads guide and the guide to running creator content as paid ads.
Before you start
- Each candidate's own record. How it was found, read and acted on, by channel, over a period long enough for its main channel to show a result, using the stages in the measuring content performance guide.
- A measurement plan for the run. The measurement plan framework asks what objective each number serves and what decision would change if it moved. Write the run's small version first: the job, the action, the event that records it, the report and who reads it.
- Your tagging convention. The source, medium, campaign and content values from the naming conventions and UTMs guide.
- An ad account and its documentation. Google Ads Help, the Meta Business Help Centre or LinkedIn Marketing Solutions Help, and the platform guides on this site, hold objectives, audiences, formats, bidding and ad policies; this tutorial does not restate them.
- A value for the action. From your own records, what the action you will pay for is usually worth: how often an enquiry becomes work, a subscriber a customer, a qualified lead a deal.
- Someone to watch the replies. Paid reach brings strangers who comment, and the organic versus paid social comparison asks that ad comments be watched as closely as post comments.
The steps
The first two steps decide whether to pay and what the run is for, the middle ones build it, and the last three set the rules, read the result and close the record.
Choose the piece from what it has already done
List the pieces you might promote and read each against its own history, not against the rest of the library. A candidate for acceleration has evidence that readers value it: search sends it engaged visits, newsletter readers click through and keep reading, it appears in journeys that end in sign-ups or enquiries. A candidate for capture answers a searched question in the page's own words yet sits outside the results people click. A candidate for a defined audience was written for a group none of your channels reaches. Set aside pieces chosen for other reasons: the newest, the most expensive, the one a colleague is proud of, the one nobody reads. Those are the pieces promotion is usually asked to rescue. If no candidate passes, that is a finding: the work lies in the pieces or the owned channels, and the budget waits.
Name the job and the one action you will judge it on
Write a sentence naming the job, acceleration, capture or a defined audience, and the action the run will be judged on. Take the action from the piece's stage, not from whatever the platform reports most readily: returning readers and assisted conversions for a consideration piece, qualified leads for a gated asset, a newsletter sign-up for a piece meant to grow the audience. Clicks and engaged visits are still read, because they explain the result, but they are not what the run is for. Add the action to the measurement plan with the event that records it and the person who reads it. A run judged on whichever figure looks best afterwards cannot fail, and so it cannot teach.
Choose the paid route that fits the job
Capture is a search job: people are already typing the question, and paid search on that query is the route the channels guide gives for it until the page ranks. Acceleration and a defined audience are usually feed jobs, where paid social puts a derivative in front of people who were not searching. For a professional audience, the LinkedIn company pages guide notes that LinkedIn's own documentation presents boosting a page post as a quick way into Campaign Manager, which suits a post already doing well organically. A colleague's post about the piece becomes a thought leader ad, with permission agreed in writing; a creator's needs the paid-media licence and platform authorisation the creator amplification guide describes. The table after these steps sets the routes side by side, and each platform's help centre covers its setup.
Define the audience from what you already know
Start from the people the piece was written for, as its brief and the audience research for content describe them, and from what its record shows about who read it and what they did next. Translate that into the targeting the platform documents, without claiming precision it does not offer. Then answer the question the prospecting versus remarketing guide frames: new people, or people who already know the business? RemarketingAdvertising to previous visitors or engagers. to past readers usually reports the cheaper result, though many of those readers would have returned without the ad. When the job is new readers, leave existing readers and customers out where the platform allows; when it is bringing readers back, say so and judge it on return visits. Read ReachUnique people who saw an ad. and FrequencyAverage times a person saw an ad. across the organic post and the paid version together, so followers are not shown the same idea several times over.
Make a derivative for the placement, not an excerpt
The promoted unit is not the article. The repurposing guide sets out what a derivative takes: an opening written for the reader in that feed or results page, a length that suits the placement, the channel's voice and a call to action chosen for it. Keep the claim, the evidence and the example; drop the exact text, because a pasted introduction under a link is scrolled past. Make a promise the page keeps: click-through (CTRShare of impressions that result in a click.) shows whether the promise suited the audience, and what readers do after the click shows whether the page lived up to it. Where the budget allows, cut more than one derivative, so the stop rule has something to choose between. Once money is behind it, a promoted post is an advertisement and its claims are advertising claims; a person's voice brings the disclosure and permission rules of the thought leader ads and creator guides. Nothing here is legal advice.
Give a cold reader a next step on the page
Readers who arrive from an ad know less than readers who arrive from your newsletter: they have not met the brand and came because of a derivative, not the headline. Read the piece as they would. Add a line of context where it assumes knowledge, check it on a phone, and make the next step suit the job: a newsletter offer for an audience-building piece, the next piece in the journey for a consideration piece, an enquiry or quote link where the piece itself says the reader will need help. The diagnosis for traffic that is not converting counts a missing next step among the commonest reasons readers leave. Do not add a form to an open piece for the sake of the run; gating is a decision about the asset, not the campaign.
Tag every link and record the run
Every paid link carries campaign parameters (UTMURL tags identifying traffic source.) built with the house convention: a source and medium that put the visits in the paid channel, a campaign value naming the piece and the run, and a content value for each derivative. The UTMs guide also explains where a platform's own auto-tagging replaces manual tags, and why untagged social ad links tend to be counted as organic social. The channels guide makes the same point: without labels on both, the paid and organic results for one piece cannot be set side by side. Then open a short run record beside the piece in the content inventory or campaign log: the piece, the job, the action, the route, the audience, each derivative with its tag, the cap, the stop rule, the dates and the owner.
Set the budget and the stop rule before launch
Set the money from your own records, never from another organisation's benchmark. Decide the most you will pay for the action the run is judged on, given what it is usually worth, and work back to the most you will pay for an engaged visit, using how often engaged readers of this piece have taken that action before. Then write the rule in full: what each derivative gets to prove itself, the cost above which it stops, the cap and end date for the whole run, the result that earns a second run, and whether money from a stopped derivative moves or stays unspent. Write it before launch, because once a run has started every figure looks as if it might improve tomorrow. Read the platform's help centre on how its budget and delivery settings work before you choose the amounts.
Read engaged visits and next steps against the owned channels
When the run ends or the rule fires, set each derivative's figures beside what the owned channels did for the same piece over the same weeks. The CTR calculator turns impressions and clicks into click-through, and the reach and frequency calculator estimates reach and average frequency from impressions and audience size where the platform does not report them. Read engagement and outcomes together, as the measuring guide asks: engaged visits, the action, the cost of each, and how many readers were new. The owned channels delivered their readers at no media cost, so the paid figure has to buy what they could not. Treat assisted conversionsConversions in which a channel appeared earlier on the path but was not the final interaction. as modelled estimates and name the model. These figures describe what happened alongside the ads, not what the ads caused; that is the question an incrementality test answers by withholding promotion from a comparable group, so for one small run write the question down rather than claim the answer.
Scale, re-cut or stop, and record why
Every run ends in one of three decisions, written into the record with its reason. Scale when a derivative met the bar on the action that matters: run it again to more of the same audience under a new cap, watching frequency. Re-cut when the pattern points at the derivative or the page: many clicks with readers leaving at once means the promise and the page disagree, while few clicks with good reading means the derivative undersold the piece. Change one thing and run again under the same rule. Stop when cold readers do not stay with the piece whatever the derivative; that sends the piece back to its brief, not to a bigger budget. Bring the record to the monthly review the measuring guide describes, and let the pattern shape the next briefs. The distribution hub sets this beside the routine every piece follows.
Matching the paid route to the job
The route follows from the job and from what the piece has already shown. The table pairs them with the measure the run is judged on and where the setup is covered.
| Job | What the piece has already shown | Paid route | Judge it on | Where the setup is covered |
|---|---|---|---|---|
| Acceleration | Readers found it, read it and acted without help | Paid social promotion of a derivative to more people like its readers | Cost per engaged visit and per action, beside the owned channels | Organic vs paid social; the Meta Business Help Centre |
| Capture | People search its question, but the page is not yet among the results | Paid search on that query until the page ranks | Cost per visit and the actions those visits assist | Distribution channels; Google Ads Help |
| A professional audience no owned channel reaches | It suits a role or industry the company page can address | Promoting a company page post that is already doing well | Engaged visits and actions from that audience | LinkedIn company pages; LinkedIn Marketing Solutions Help |
| A colleague's view of the piece | An employee's post about it earns attention unpaid | A thought leader ad, with permission in writing and any connection disclosed | Its results compared with the brand's own ads in that account | Thought leader ads |
| A creator's coverage of the piece | A creator's post about it already performs | The creator's post run as an ad, under a paid-media licence and platform authorisation | The action the run was set up to judge | Amplifying creator content |
| Bringing readers back | Its readers go on to deeper pieces when prompted | Remarketing to recent readers with the next piece in the journey | Return visits and assisted conversions, with a holdout where possible | Prospecting vs remarketing; content for consideration |
| A gated asset | Buyers trade contact details for it and sales can act on them | Paid promotion of the asset to a defined audience | Qualified leads per asset, with production counted in the cost | Content for lead generation; gated vs ungated content |
One piece can justify more than one row: a guide that ranks may earn acceleration on social while a newer page on the same subject is supported with paid search. Keep them as separate runs with their own tags, caps and stop rules, so a strong result on one route does not hide a weak one on the other.
A furniture restorer promotes its finishing guide (illustrative)
Zibrovenk Restoration promotes the guide readers already act on
Zibrovenk Restoration, an invented furniture restoration workshop, publishes guides its owner writes herself. Last quarter its guide to choosing a wood finish ranked for its question, drew 2,300 visits from search, 1,320 of them engaged, and appeared in the journeys of 44 of the workshop's 63 quote requests. Its guide to repairing lifted veneer, published five months ago, has 150 visits and appears in none. Her instinct is to promote the veneer guide because it needs readers; the record points to the finish guide because readers already act on it. Her job sentence: accelerate the finish guide to people within 20 miles, judged on quote requests. About one engaged reader in 30 asks for a quote, and she will pay up to £36 for a quote request, so an engaged visit is worth up to £1.20. Instead of posting the link, she cuts a 30-second before-and-after clip of a table being refinished and a five-frame carousel comparing finishes, adds a quote link where the guide says a piece needs a professional, and tags each link with paid social as the medium and clip or carousel as the content value. The written rule: a £250 cap over three weeks split evenly between the two derivatives, £60 for each to prove itself, any derivative above £1.20 per engaged visit at that point stops, money from a stopped derivative stays unspent, and a derivative that brings quote requests at £36 or less each earns a second run. After £60 the carousel has reached 3,900 people at a frequency of 1.8, with 62 clicks (0.9%), 38 engaged visits at £1.58 each and no quote request; it stops. The clip spends its £125, reaching 8,600 people at a frequency of 2.2, with 280 clicks (1.5%), 170 engaged visits at £0.74 each and 5 quote requests at £25 each, 4 from people with no earlier visit. Over the same weeks the newsletter sends 140 engaged visits and 3 quote requests at no media cost, all from existing subscribers. The record reads: clip to a second run under a new cap and the same thresholds, carousel stopped by the rule at £1.58 per engaged visit (few clicks, but readers who clicked stayed, so a re-cut is the next test), £65 unspent, and whether the 5 requests would have come anyway unanswered. The veneer guide goes back to its brief.
Common mistakes
- Promoting the piece that needs help. Paying multiplies what a piece already does; the weakest piece in the library is the least likely to repay it.
- Boosting the link instead of making a derivative. A pasted title and introduction is an excerpt in a feed built for native posts.
- Judging the run on clicks or reach. Click-through says the promise suited the audience; only engaged visits and the action say whether the piece kept it.
- Writing the stop rule after launch. A rule set once the figures are in bends to fit them.
- Untagged paid links. Paid visits blur into organic and direct traffic, and the comparison with the owned channels cannot be made.
- Showing it mainly to people who already read it. Past readers report the cheapest result, much of which would have happened anyway.
- Borrowing someone else's numbers. Costs and rates from another business or market say nothing about yours; the ceiling comes from what the action is worth to you.
- Reporting the attributed figure as lift. Actions recorded beside the ads are not proof the ads caused them; name the model and leave causation to a test.
