The short answer
Retargeting should continue the conversation
Most retargeting is set up backwards. An agency creates one audience called “all website visitors”, puts every old advert in it and reports a cheaper click-through rate. The numbers look warmer because the people are warmer. Nothing has been learned.
The commercial version asks a harder question: what does this action tell us, and what would genuinely help that buyer move one step closer?
The lazy version
Everyone visits something
Everyone enters one audience
Everyone sees “Book a demo”
The useful version
Separate meaningful signals
Choose the missing argument
Exclude people who moved on
The audience sources
Your website is only one source of warm attention
This matters for a niche firm. If the site gets a few hundred suitable visits a month, splitting those visitors into seven tiny audiences will leave none of them large enough to run. LinkedIn activity can add useful signals without pretending weak engagement is buying intent.
Website actions
Insight Tag required
Visited a page, clicked or submitted on your site
Best when the page tells you something useful about intent.
Video viewers
Built inside LinkedIn
Watched 25%, 50%, 75% or 97% of a video ad
Useful for moving an engaged viewer to a demo, guide or proof.
Lead Gen Forms
Built inside LinkedIn
Opened or submitted a LinkedIn form
Separate the curious non-submitters from people who already converted.
LinkedIn engagement
Built inside LinkedIn
Engaged with an ad, Company Page or LinkedIn Event
Useful extra volume when a niche website audience stays too small.
The message map
Show the next argument, not the last advert again
The funnel is not a neat staircase, but the previous action still gives you context. Use it. This table is a planning tool, not a claim that every visitor has the same intent.
| Signal | What it might mean | Useful next message | Avoid |
|---|---|---|---|
| Read one useful article | Early interest | Another strong idea on the same costly problem | A hard meeting request |
| Watched most of a demo | Product interest | A customer example, technical answer or fuller demonstration | Repeating the same video |
| Opened a form, did not submit | Considered the offer | Remove the likely doubt or give a lower-friction route | Pretending they never saw it |
| Visited pricing or contact | Commercial interest | Scope, process, proof and a clear route to talk | Generic awareness copy |
| Known stalled opportunity | Commercial history | Answer the reason it stalled or show what has changed | A creepy 'we saw you' message |
Your warm audience should not be seeing leftover adverts
We can map the signals, build the audiences and create the next messages around your real buying cycle.
The niche B2B problem
What to do when the audience will not clear 300
LinkedIn requires at least 300 matched member accounts before an ad set can deliver. That is a privacy floor, not a recommended campaign size. A source can show hundreds of actions and still match fewer than 300 LinkedIn members.
Traffic
People arrive
Match
LinkedIn recognises some
Filters
Location narrows it
300+
Delivery can start
Widen the lookback window
A 30-day website pool may fail where 90 or 180 days works. Use a window that still fits the sales cycle.
Combine related pages
Group pages that show the same type of interest. Do not mix everything merely to force delivery.
Use LinkedIn engagement
Video viewers, form opens and Company Page engagement can build without relying on website cookies.
Build the audience deliberately
Run useful video or document ads to your named market, then retarget the people who paid attention.
Use named companies
If you know the firms you want, upload the company list and segment by the job titles involved in the purchase.
Accept when LinkedIn is not viable
If the real audience stays too small, use Meta, email, sales follow-up or search instead of padding it with bad-fit people.
For the named-company route, read how to upload a target account list and layer the buying roles.
A usable account structure
Keep the campaigns separate enough to learn
Do not create a campaign for every page on the site. Do separate signals when the audience size and next message justify it. A sensible starting structure for many niche firms looks like this.
This is a starting structure, not a platform rule. Combine rows when audience size is too low to make the separation useful.
The adverts
Warm does not mean ready
Somebody can read an article because the headline caught them. A video view can be professional curiosity. Treating every engagement as a buying signal produces months of desperate “still interested?” adverts.
For early attention
Make the problem clearer
A founder explains the expensive workaround buyers have normalised.
For product interest
Show the work
An engineer gives the best sales demonstration once, then the right market can watch it all year.
For consideration
Remove the risk
A real case study, implementation walkthrough or answer to the question sales usually gets late.
For internal justification
Make the cost visible
A calculator or comparison that helps the buyer price the current problem.
For lapsed pipeline
Deal with the objection
Turn the reason a real opportunity stalled into a useful advert or article.
For high intent
Ask directly
Clear scope and a straightforward booking route when the prior signal supports it.
Need a better demonstration?
Turn the best sales pitch in the business into a video that keeps working.
Use the B2B video ad blueprintWindows and exclusions
Recency matters, but so does the buying cycle
LinkedIn supports different lookback windows depending on the audience source. Choose them around how quickly the signal loses meaning. A pricing-page visit from last week and an article read eleven months ago should not carry equal weight.
Use windows as hypotheses
- Shorter for contact, pricing and live buying signals.
- Longer for educational content and long sales cycles.
- Separate recent from older interest only when both groups can deliver.
- Review whether old signals still deserve the same advert.
Exclude by default
- People who completed the current action
- Existing customers where the offer is irrelevant
- Employees and direct competitors
- Known poor-fit companies
- Overlapping warmer audiences from broader campaigns
Keep Audience Expansion off. If LinkedIn can move beyond the people you deliberately selected, you no longer have a clean retargeting campaign.
The wear-out problem
A small warm audience can turn expensive very quickly
Retargeting often looks efficient at launch because the audience already recognises the problem. Then frequency climbs, the same people see the same creative and response falls away. More budget makes that happen faster.
01
Useful familiarity
The buyer recognises the name and idea.
02
Repeated proof
Different content strengthens the case.
03
Creative fatigue
The same message stops earning attention.
04
Irritation
More delivery now makes the brand easier to ignore.
Watch frequency alongside response over the same date range. If delivery rises while clicks, useful visits and conversions flatten, refresh the message or reduce spend. Do not celebrate cheap impressions that nobody wants.
Retargeting should make the rest of your marketing work harder
We build the warm-audience system around your traffic, content, sales cycle and CRM rather than treating it as a leftover campaign.
UK privacy basics
The Insight Tag should not quietly fire before consent
Website retargeting uses an advertising tag. UK guidance on cookies and similar technologies normally requires clear information and active consent before advertising technology stores or accesses information on a visitor's device. A privacy-policy mention on its own is not the same thing.
A sensible working checklist
- Put the Insight Tag behind the advertising-cookie choice in your consent system.
- Explain LinkedIn's role in the cookie and privacy information.
- Test that rejecting advertising cookies prevents the tag from firing.
- Do not place the tag on pages collecting sensitive consumer information.
- Check the current ICO and LinkedIn guidance for your own market and setup.
This is practical advertising guidance, not legal advice. Consent rules and your lawful basis depend on the implementation and the people you target.
The scorecard
Judge whether warm attention becomes commercial movement
A higher click-through rate does not prove retargeting worked. Warm audiences often click more because they already know the business. The useful report shows whether the campaign moved the right companies towards a qualified conversation.
Direct
The enquiry has a traceable LinkedIn source. Report it separately.
Influenced
The account saw or engaged with ads before moving through another route. Show the evidence without claiming all the credit.
See the full method for measuring LinkedIn influence and matching company engagement against CRM pipeline.
The human follow-up
Use the signal without making the prospect feel watched
Retargeting can tell marketing that an audience or account is warming. It should not give sales a reason to write, “I saw you visited our pricing page.” That is not personalisation. It is surveillance with a greeting.
Do not say
“We noticed you were looking at our website, so I wanted to follow up.”
Use the context instead
Contact the right account with the same useful idea, question or proof that earned attention in the campaign.
This is strongest when the CRM holds audience membership, campaign history and deal stage at account level. Marketing can see which messages create attention. Sales can say whether that attention became a useful conversation. Together they decide what the account sees next.
If the programme starts from a fixed company list, use the account-based marketing guide. For the wider system around it, read the B2B demand generation playbook.
First-party case study
The budget did not grow. The follow-up changed.
For a large IT company, 10% of the existing advertising budget was moved into LinkedIn retargeting. The campaign used video to keep the business visible to people who had already visited. Enquiry calls doubled. The full case study shows the setup and its limits.
10%
of existing budget reallocated
2×
enquiry calls in this case
One client result, not a forecast or general benchmark.
Read the case studyThe first month
Build the useful audiences before asking them to perform
Week 1
Check
Verify the Insight Tag, consent behaviour and conversion actions. List every usable LinkedIn engagement source.
Week 2
Build
Create audiences by signal, set sensible windows and exclusions, then see which can clear the minimum.
Week 3
Make
Create the missing next message for each viable audience. Do not recycle cold adverts by default.
Week 4
Launch
Start cautiously, watch frequency and response, then connect qualified actions to the CRM.
The failure points
Where LinkedIn retargeting goes wrong
Everything goes into one pool
An article reader and a contact-page visitor receive the same hard sell.
The audience never reaches 300
The campaign sits ready forever because the source, window or location filter is too narrow.
The same advert follows people
Retargeting repeats the introduction instead of adding a reason to believe.
Converted leads stay included
Budget keeps chasing people who already did the thing the advert asks for.
Audience Expansion stays on
LinkedIn moves beyond the warm pool and the reporting no longer answers the same question.
Budget ignores audience size
A small list is forced to absorb spend until frequency becomes the only number rising.
Consent is treated as a banner
The tag fires before a choice, so the words and the behaviour do not match.
Warm clicks become fake revenue
Every later deal is credited to retargeting even when the evidence only shows prior exposure.
The question to take into Monday
What should a buyer see after they pay attention?
If the answer is “the same advert again”, start there.
Retargeting is built into every account under my LinkedIn Ads management for B2B, alongside the named-account campaigns it follows up.
Common questions
Evergreen LinkedIn retargeting
Turn paid attention into a better next move
We connect the audiences, creative, exclusions, landing pages and CRM evidence. You see what the warm campaign reached, what people did next and whether the right accounts moved.