How to spend a $10,000 LinkedIn ads budget without wasting it

Allocate $10,000 in reverse: retargeting cap first, test pot second, cold reach last

Build the plan backwards from the two numbers that are computable, and let cold reach absorb the remainder. Most plans do the opposite. They start with cold reach because it is the biggest line and the most enjoyable to argue about, then treat retargeting and testing as leftovers. The result is a cold campaign that is too wide, a retargeting campaign that hammers the same forty people, and a test that never reaches a sample worth reading.

Two of the three lines can be calculated from inputs you already have in the account. Only the third involves judgment, and by the time you reach it the judgment is constrained by arithmetic. That constraint is the entire benefit.

Worked on an assumed CPM of $35 and an assumed click-through rate of 0.45%. Replace both with your own account numbers before you use this.

Look at the size of the retargeting line. Five hundred and sixty dollars against a pool of two thousand people is not a rounding error, it is the correct amount. Spending three thousand there would buy the same faces more times than any human being can tolerate.

Retargeting is a capped number, not a percentage of the budget

The retargeting budget has a ceiling and the ceiling is set by how many people are in the pool. Above it you are not reaching more people, you are reaching the same people more often, and past a certain frequency the extra impressions actively cost you goodwill rather than merely wasting money.

The formula is pool size multiplied by the frequency you want per month, multiplied by CPM, divided by one thousand. Everything in it is a number you can read off a screen except the frequency, which you choose.

Monthly retargeting ceiling at a target frequency of eight and an assumed CPM of $35.

Frequency of eight per month is a working assumption and deliberately on the generous side, because a retargeting pool is by definition people who already raised a hand. Pick your own number and defend it. The argument that matters is not which frequency is correct, it is that a ceiling exists and that most retargeting budgets are set without anyone calculating it.

The Starve Line tells you how many creatives $10,000 actually supports

A creative that never receives enough impressions to accumulate a readable number of clicks has not been tested, it has been sampled. The Starve Line is the minimum spend one creative needs per flight, and dividing your test pot by it gives the honest creative count. It is almost always lower than the number in the plan.

Starve Line sensitivity. The creative count changes by a factor of two across a click-through rate range that is entirely plausible.

The one specific figure circulating in account based marketing writing, ten to fifteen ads per persona at eight to ten thousand dollars, is not wrong. It describes a rotation rather than a test. Fifteen creatives sharing a $2,000 test pot gives each one about $133, which at any click-through rate above returns a handful of clicks. Run fifteen if you want variety in the feed and you have accepted you will never rank them. The mistake is running fifteen and then acting on the ranking.

Cold reach takes the remainder, and the remainder sets the maximum audience size

Once retargeting and testing are funded, the cold budget is fixed, and a fixed cold budget imposes a hard maximum on audience size. Most $10,000 plans get this backwards and pick an audience of two hundred thousand people on the grounds that two hundred thousand people exist.

Here is the arithmetic on the running example. A cold budget of $7,440 at an assumed $35 CPM buys about 212,600 impressions for the month. Divide by the audience to get impressions per person, then hold that against a floor of three.

Impressions per person from a $7,440 cold budget at an assumed $35 CPM.

The cut is usually geography or seniority, and the internal argument you will have is that cutting the audience reduces the opportunity. Arithmetically it increases the opportunity, because an audience nobody saw twice was never an opportunity in the first place. Say it in impressions per person and the argument shortens considerably.

A four week ramp that does not reset your delivery

Raise the budget by less than half at a time and hold each step for a full week. The reason is pacing behaviour, not superstition, and understanding the reason tells you when the rule can be broken.

A campaign paces its delivery against the budget it currently has. A large jump pushes the system to buy inventory it had previously been outbidding, which lifts the average cost per click at precisely the moment you are trying to prove the channel works. LinkedIn does not publish a threshold for this and you will not find one in the help centre. The rule practitioners hold to is under fifty percent per change, and it is the mechanism that makes it sensible rather than any published figure.

A four week ramp totalling $10,000. Every step is under a fifty percent increase on the week before it.

If your $10,000 is a one-off rather than a monthly commitment, run the same four weeks and stop. Do not stretch it across eight weeks at half the daily rate. That halves density in every single week and leaves you with eight weeks of unreadable data instead of four weeks of a decision.

What to cut first when the month is half gone and the numbers are bad

Cut in the reverse order of the allocation. Testing goes first, cold audience width goes second, retargeting goes last, and the offer only gets touched when the click numbers say it is the offer. Cutting in a different order is how a recoverable month becomes a cancelled channel.

Day fourteen decisions. Two of these rows look identical on a dashboard summary and require opposite responses.

Before cutting anything, route the failure to the broken number, because rows two and three above produce very similar summary screens and opposite instructions. And if the complaint is lead quality rather than volume, junk leads have a separate cure and most of it sits upstream of the ad account entirely.

The reporting lines that decide whether month two gets the same $10,000

Report cost per qualified meeting and pipeline created first, and show cost per lead fourth. A finance conversation that opens with cost per lead ends with a budget cut, because cost per lead improves fastest when lead quality falls, and everybody in the room eventually works that out.

1. Qualified meetings booked, using a definition sales agreed in writing before the month started.
2. Cost per qualified meeting, which is spend divided by that number and nothing else.
3. Pipeline created, attributed with a stated model, including a plain sentence about where the model is weak.
4. Cost per lead, shown fourth, with a note on what changed in lead quality alongside it.

Agreeing the definition of qualified before the month begins is the whole trick. Agreeing it afterwards turns every review into an argument about definitions, and the ads lose that argument every time regardless of what they produced.

The distance between the first two bars is the subject of this article. Almost everyone runs the channel for lead generation and a much smaller group reports that it produces leads for them. The difference is rarely the platform and usually the allocation.

Where $10,000 disappears even when every setting is correct

Three leaks account for most of it: audience overlap between your own campaigns, slow follow-up, and an offer shaped for the seller. None of the three shows up as a red flag in Campaign Manager, which is why they survive month after month.

Audience overlap, where you outbid yourself

Two campaigns aimed at overlapping audiences meet in the same auction, and you are both the underbidder and the overbidder in a single transaction. Exclude every campaign's audience from the others, and specifically exclude the retargeting pool from the cold campaign, otherwise your most expensive impressions land on people you were already reaching cheaply.

Follow-up latency, which no bid strategy can fix

A form fill that sits untouched for two days has cooled, and no amount of bidding sophistication recovers it. The fix is a named owner and an agreed response time, both of which live outside the ad account. If nobody owns the response, reduce the budget until somebody does.

An offer shaped for the seller

A demo request suits the seller. At $10,000 a month you can afford an offer that suits the buyer at the stage they are actually at, which usually means something they can use before they talk to anyone. The conversion difference is normally larger than anything creative testing will find.

There is also the question of what the click lands on and who appears to be talking. Whether the CEO or the company page should carry the message changes how a cold impression is received, and a paid budget pointed at a page with nothing on it is a slower version of the same leak.