Budget & bidding — how the auction works
It is an auction: highest bid × relevance score wins and pays just enough to beat the runner-up. Better creative buys cheaper media.
The auction in one line
Winner = highest (bid × relevance score), and pays just enough to beat the runner-up (second-price logic). Relevance score = predicted engagement/CTR of the ad for that member. Better creative lowers the price.
| Strategy | What it does | When |
|---|---|---|
| Maximum delivery (automated) | Spends full budget at whatever price wins | Default; large audiences, deadline campaigns, no time to manage |
| Cost cap | Targets an average cost per key result; can under-deliver | Efficiency with a known target CPC/CPL |
| Manual bidding | You set the max bid per click/impression/view/send | Small audiences, experienced buyers; start under the suggested range, +10% steps |
| Qualified Leads Optimisation (QLO) | Send MQL/SQL signals back via CAPI; delivery skews to similar members | ≥ ~50 qualified events / month |
| Conversion Value Optimisation (CVO) | Optimises toward higher-value conversions | Value passed via CAPI |
How the second-price auction settles
LinkedIn runs a modified second-price auction: the ranking uses bid × relevance, and the winner pays the lowest bid that would still have won, not the bid it entered. Every time a member opens the feed, this happens in milliseconds:
- Eligibility. Every ad set whose targeting matches the member, that still has budget and that pacing allows to bid right now, enters.
- Prediction. LinkedIn predicts how likely this member is to engage with each ad: the relevance score.
- Floor check. Bids below the floor for this audience and placement are dropped.
- Ranking. Score = bid × relevance. Highest score wins the slot.
- Price. Winner pays runner-up's score ÷ winner's relevance + €0.01. Never more than the bid, never less than the floor.
- Billing. You are charged in the ad set's unit: per impression, click, view or send.
| Advertiser | Bid | Relevance | Score | Rank | Pays |
|---|---|---|---|---|---|
| B — great creative | €8 | 1.00 | 8.0 | 1st | 7.0 ÷ 1.00 + 0.01 = €7.01 |
| C | €10 | 0.70 | 7.0 | 2nd | nothing — lost |
| A — dull creative | €12 | 0.55 | 6.6 | 3rd | nothing — lost |
Bid floors and pacing
Floors — the reserve price
- Every audience and placement has a floor (minimum bid). With manual bidding, Campaign Manager shows it and will not save a bid below it.
- If you are the only bidder above the floor, you pay the floor, not your bid.
- The floor rises with demand: senior titles, small high-value audiences, IT, finance and HR in DACH, quarter-ends and Q4.
- Suggested bid range ≠ floor. It is a forecast of what competing advertisers pay for this audience. The site's rule still holds: start under it, then raise in +10% steps.
- Text ads have a fixed minimum bid of $2 CPC or $2 CPM.
Pacing — how the money is spread
- LinkedIn paces automatically. A daily budget is spread across the hours when your audience is active, so €300 does not disappear at 09:00. A lifetime budget is spread across the flight.
- With Maximum delivery, pacing and the bid are automatic: the system bids whatever it needs to use the budget.
- With manual bids or a cost cap, pacing tries to spend but the win rate limits it. If you lose most auctions, you see it as underspend. The money is not saved; the plan simply does not happen.
- A total budget and schedule on the Campaign (formerly Campaign Group) caps all its ad sets together. Use it when the IO is a fixed amount.
- Rule of thumb, not a LinkedIn rule: change budgets and bids in 20–30% steps, then wait 2–3 days before judging. Big jumps reset what automated bidding has learned.
Manual vs maximum delivery vs target cost vs cost cap
| Strategy | You set | LinkedIn does | Spends full budget? | Main risk | Use when |
|---|---|---|---|---|---|
| Maximum delivery | Budget only | Bids automatically to get the most results for the budget | Yes, almost always | Cost per result drifts up on small or competitive audiences | New account, first 2–3 weeks to learn the market price, deadline campaigns, awareness |
| Target cost (legacy) | A target average cost per result | Keeps the average near the target; single results above or below | Usually | Average can settle above the target | Old ad sets; know it for the exam |
| Cost cap | A ceiling on the average cost per key result | Maximises results while keeping the average at or below the cap | Only if the cap is at or above market | Underdelivery | You know the CPL/CPA you can afford and have history to set it |
| Manual bidding | Max bid per click, impression, view or send | Enters auctions at up to your bid; you still pay second price | Only if the bid wins enough auctions | Too low = no spend; too high = overpaying on small audiences | Small or ABM audiences, experienced buyers, when Max delivery overpays |
A sequence that works in most accounts: launch on Maximum delivery → after 2–3 weeks read the real cost per result → switch to a cost cap 10–20% above that number, or to a manual bid just under the average CPC/CPM paid → tighten in small steps.
Cost cap behaviour
- The cap applies to the average cost per result over the ad set's life, not to each result. Single leads can cost more than the cap.
- You are still billed per impression or click. The cap is what the system steers the average cost per result (lead, conversion, click) towards.
- A cap below what the market costs does not overspend, it underspends. Delivery slows until the auction gives results at that price, or never does.
- The first days are often above the cap while the system learns. Judge after 7 days, not 2.
- Raising the cap by 10–20% opens more auctions. Improving creative (relevance) wins more of them at the same cap. That is the cheaper fix.
- A cap needs budget room. Rule of thumb: a daily budget of at least 3–5× the cap, or the system sees too few results to steer.
- Do not stack a tight cap on a tiny audience with Audience Expansion off. That is the classic €0-spend combination.
Pricing units
Budgets, pacing, frequency
- Minimum $10/day per ad set, $100 lifetime (EUR equivalents shown in account currency).
- Daily budget can be exceeded by up to 50% on a given day; LinkedIn averages over the calendar week — say this before finance asks.
- Rule of thumb: €1–2 / day per 1,000 members in a cold audience.
- Lifetime budgets pace across the schedule; daily + lifetime combined gives a hard cap.
- Sponsored Messaging has strict per-member frequency caps, so a member rarely sees your message twice in a short period.
- Ad review takes up to 24 h; rejections: superlatives ("No. 1", "guaranteed"), before/after images, landing page ≠ ad.
- Run 3–5 ads per ad set; refresh creative every 2 weeks on audiences < 10k, monthly > 100k.
Why an ad set does not spend — checklist in order
- Bid below the floor for the audience
- Ad in review or rejected
- Campaign (group) paused or schedule ended
- Audience too small for the format
- Daily budget below the bid
- Billing failed
- Conversion-optimised ad set with no conversion volume
- Frequency exhausted — new creative resets it
Daily vs lifetime budgets
| Daily budget | Lifetime budget | Daily + lifetime together | |
|---|---|---|---|
| What it limits | Average spend per day (+50% on a busy day) | Total spend over the flight | Both: steady pacing with a hard total |
| End date | Optional; runs until paused | Required | Required |
| Pacing | Even day by day | LinkedIn shifts spend to days with more opportunity | Lifetime pacing, but never above the daily limit |
| Best for | Always-on programmes, monthly IOs, tests | Fixed-amount flights: events, launches, trade fairs | Finance-sensitive clients: "never more than X, never more than Y a day" |
| Watch out | A +50% day surprises finance; say it upfront | Editing it mid-flight re-paces the rest; spend can bunch near the end | A daily limit that is too low means the lifetime total is never reached |
Minimum spend rules
| Rule | Value | What it means in practice |
|---|---|---|
| Daily budget per ad set | $10 (account-currency equivalent) | Floor to go live, not a useful budget; see the €1–2/day per 1,000 members rule |
| Lifetime budget per ad set | $100 | A lifetime budget also needs an end date |
| Daily overspend | up to +50% on a given day | Averaged over the calendar week, so the plan holds |
| Manual bid | ≥ the floor shown in Campaign Manager | Floors differ by audience, format and placement |
| Text ads | minimum bid $2 CPC or $2 CPM | The cheapest format to enter, and the weakest placement |
| Audience | 300 matched members to serve; 1,000+ stable | Below that, no amount of budget delivers |
| Conversion-optimised bidding | ~50 conversions / month; QLO ~50 qualified events | Needed monthly budget ≈ 50 × expected CPA |
| Useful test | ≈ €4k / month media | Less than that and the full-funnel read is noise |
| Currency | set once per ad account | Wrong currency = new ad account |
Why a campaign underdelivers
Underdelivery is a price signal: the market is telling you the bid, the cap or the creative is below what this audience costs. Go through the checklist above for "no spend at all". Use this table when it spends some, but not enough:
| Symptom | Likely cause | Fix |
|---|---|---|
| €0 from day one | Bid below floor, ad in review or rejected, billing, schedule not started | Check the Status and Delivery columns; move the bid up to the floor |
| Spends 20–60% every day | Manual bid or cost cap below market | +10% bid steps, or raise the cap 10–20%, or improve creative |
| Spent well, then fades | Frequency > 5, CTR falling, so relevance and win rate fall too | New creative, or widen the audience; lower the daily budget on small audiences |
| Spends only on some days | Small audience; competition spikes (quarter-end, Q4, trade fairs) | Widen targeting or lengthen the flight; accept that a lifetime budget will bunch |
| Conversion ad set barely spends | Too few conversions for the model to learn | Optimise for a higher-funnel event, or merge ad sets to pool volume |
| Penetration < 30% after a month | Bid too low for this audience | Raise the bid or switch to Maximum delivery for two weeks to read the price |
| Much less than forecast | Layered targeting plus Audience Expansion and Audience Network off shrank the reachable pool | Check the forecast after every targeting change; keep feed audiences above 50k where you can |
Worked examples
1 · Settling an auction, and what bad creative costs
Same three advertisers as the diagram. B wins with 8.0 and pays 7.0 ÷ 1.00 + 0.01 = €7.01.
Now B's creative gets tired and relevance drops to 0.80: score 8 × 0.80 = 6.4. That is below C (7.0) and A (6.6), so B is third and loses. C now wins and pays A's score ÷ C's relevance = 6.6 ÷ 0.70 + 0.01 = €9.44, still under its €10 bid.
Add a floor of €7.50 per click for this audience. If B still had relevance 1.00, it would pay max(€7.01, €7.50) = €7.50.
Lesson: the same €8 bid went from a €7.01 click to no click at all. Only the creative changed.
2 · Sizing a budget from the audience
Audience: 70,000 IT managers in DACH at 200–1,000-employee companies. Rule: €1–2 per day per 1,000 members → €70–140 / day.
- Pick €100/day → ≈ €3,040 / month (30.4 days).
- At a €50 CPM: 100 ÷ 50 × 1,000 = 2,000 impressions / day → ≈ 60,800 / month.
- If 50% of the audience is reached (35,000 members), frequency = 60,800 ÷ 35,000 ≈ 1.7 per month. That is too thin for recall, so either narrow to the best 35k or go to €140/day.
3 · A manual bid ladder
Suggested CPC range €11–16, daily budget €100. Start under the range at €9.50.
| Day | Bid | Spend of €100 | Avg CPC | Action |
|---|---|---|---|---|
| 1–3 | €9.50 | €22 | €9.30 | Winning too little → +10% |
| 4–6 | €10.45 | €61 | €9.85 | Improving → hold |
| 7–9 | €10.45 | €96 | €10.10 | Full spend → lock |
The same audience on Maximum delivery had averaged €13.40. Manual saves (13.40 − 10.10) ÷ 13.40 ≈ 25% per click, in line with the study finding that bidding low cuts cost per result by about a third.
4 · A cost cap that underdelivers, and three fixes
Client can afford a €160 CPL. Cost cap €160, daily budget €250 (€1,750 / week), audience 25k.
Week 1: spent €480 (27%), 3 leads, CPL €160. The cap holds, but the market price for this audience is nearer €200.
- Raise the cap 15% to €185: more auctions open; CPL rises toward the cap.
- Widen the audience 25k → 60k: more cheap auctions to win at €160.
- Better creative, e.g. a document ad (≈42% lower CPL): relevance rises, so the same cap wins more.
Week 3, with the document ad and the cap unchanged: spent €1,520, 10 leads, CPL €152. The creative fixed the delivery, not the money.
5 · Daily vs lifetime for a 21-day trade-fair push
Budget €6,300 over 21 days; the fair is at the end of week 3.
- Daily €300: steady; any day may spend up to €450, but the week averages €2,100. Total ≈ €6,300.
- Lifetime €6,300 + end date: LinkedIn spreads it, e.g. €200 on weekends and €380 midweek, and finishes on budget.
- You want the weight in the final week: lifetime pacing will not do that. Use a daily budget and change it: €200 → €200 → €500 per day = 1,400 + 1,400 + 3,500 = €6,300.
6 · Can this budget feed automated conversion bidding?
Website conversions (demo requests), expected CPA ≈ €280. Needs ~50 conversions a month → 50 × 280 = €14,000 / month.
Client budget €5,000 → 5,000 ÷ 280 ≈ 18 conversions, too thin to learn. Options: optimise for a Lead Gen Form content lead (≈ €150 → 33 / month), pool ad sets into one, or run Website visits plus retargeting and let the conversions come from the retargeting pool.
7 · Why a small ABM ad set stalls in week 2
Matched audience 12,000, daily €200, CPM €70 → 200 ÷ 70 × 1,000 ≈ 2,857 impressions / day → 40,000 in two weeks.
Penetration 55% = 6,600 members reached → frequency 40,000 ÷ 6,600 ≈ 6.1 in two weeks. That is above 5, so CTR falls, relevance falls, the price rises and spend stalls.
Fix: halve the daily budget to €100, or add 3 new ads, or add adjacent job functions to the list.
- Maximum delivery spends everything; Cost cap targets an average cost; Manual sets your own bid.
- CPM for awareness, CPC for clicks and leads, CPV for video, CPS for messaging.
- Daily budget is averaged over the calendar week — say it before finance asks.
- The winner pays the runner-up's score ÷ its own relevance + €0.01. Never more than the bid, never less than the floor.
- Cost cap caps the average. A cap below market does not overspend, it underspends.
- Underdelivery = bid, cap or creative below market. Automated conversion bidding needs ≈ 50 × CPA a month.