SaaS PPC budget allocation

Most paid budgets get set in a fifteen minute conversation that starts with last year’s number and a percentage. That produces a figure nobody can defend in June when the CFO asks what the extra $30,000 a month bought.

The alternative takes about an hour and produces a number with a chain of reasoning attached. Start at the ARR you need, walk backwards through every conversion rate between there and a click, and the budget falls out of the arithmetic. Then check it against the industry percentage, rather than starting there.

Work backwards from the ARR target, not from a percentage of revenue

Take the new ARR you need paid media to produce, divide it by ACV, then divide backwards through every stage rate until you reach leads, and multiply by cost per lead. That is the whole method.

Here it is with real numbers for a company targeting $2M in new ARR from paid, at a $30,000 ACV.

Two things fall out of that table immediately. The implied marketing CAC is $11,300 per customer, which you can hold up against your ACV and gross margin without any further work. And if you only have $30,000 a month, the honest conclusion is that your paid target should be $950,000, not $2M, and somebody needs to hear that in October rather than in July.

Where you have no historical rates, borrow typical ones and mark them clearly as assumptions. Our SaaS PPC benchmarks carry stage rates by ACV band and motion, which is closer than a single blended figure from an agency blog.

What the percentage of revenue rules are actually good for

They are a sanity check, not a plan. SaaS Capital puts median marketing spend for private B2B SaaS near 8 percent of ARR, and finds venture backed companies spend roughly 58 percent more as a share of revenue than bootstrapped ones. Paid media is typically a quarter to a half of that total.

Notice the $1M row. Eight percent of $1M funds a paid budget that sits below the learning floor of a single channel, which is the arithmetic reason most seed stage paid programs fail. It is not that the team was bad at Google Ads. They were running a channel at a spend level where the channel cannot function.

Venture backing changes the answer, and that is the point of the SaaS Capital finding rather than a footnote to it. A funded $1M ARR company spending 25 percent of revenue on marketing has a real paid budget. A bootstrapped one at 8 percent does not, and should be spending that money on organic and review site presence instead.

The learning floor: what each channel costs before it works

Every channel has a minimum monthly spend below which its algorithm cannot learn and your data cannot reach significance. Spend below the floor and you are not running a small campaign, you are running a broken one.

Channel	Practical monthly floor	Why the floor exists	Time to first reliable read
Google Search	$3,000	Smart Bidding needs roughly 30 conversions in 30 days per campaign	4 to 6 weeks
LinkedIn Ads	$5,000	High CPMs plus a 7 to 10 day optimisation window per campaign group	6 to 8 weeks
Meta retargeting	$1,500	Audience size and frequency caps below that make delivery erratic	3 to 4 weeks
Reddit Ads	$2,500	Subreddit level delivery fragments spend across small audiences	5 to 7 weeks
Capterra	$500 minimum plus a $2 minimum CPC	Platform contract minimums, not algorithmic	4 weeks
G2 placement	About $25,000 a year	Annual contract minimums on category sponsorship	One quarter, contractually locked

The Google Search floor deserves the most attention because it is the one people argue with. Thirty conversions in thirty days per campaign is the published guidance for target CPA bidding. At a $180 CPL that is $5,400 a month in one campaign. You can get under that with manual CPC and a very tight keyword set, which is exactly what I would do at $3,000, and it is also why $3,000 accounts need a human in them weekly while $30,000 accounts do not.

G2 is the outlier because the floor is contractual rather than technical. You cannot run a $2,000 test. You sign for a year, which makes it the one channel where the decision has to be made on modelling rather than on a pilot.

Allocation templates at $8K, $40K and $150K a month

These are starting splits for a sales-led B2B SaaS product with an ACV between $15,000 and $60,000. Adjust the search share down and the LinkedIn share up as ACV rises, because search volume runs out long before your addressable market does.

The seed column is the controversial one and I will defend it. At $8,000 a month you get one real channel plus retargeting plus a small experiment line. Search is the right single channel for almost every B2B SaaS product because it captures demand that already exists, reads fastest, and can be turned off on a Tuesday. LinkedIn at $2,000 a month is a donation.

The Series A column adds LinkedIn above its floor and buys into review sites, which is usually the moment those become worth it because you now have enough reviews for the traffic to convert. The Series B column is the first one that can honestly fund experiments at a scale where they produce answers rather than anecdotes.

Brand search stays small in all three columns on purpose. It is the cheapest lead you will ever buy and the least incremental one, since most of those people were already coming. Fund it as insurance against competitors bidding on your name, not as a growth line. Our SaaS PPC mistakes breakdown covers how brand spend quietly inflates a program’s reported performance.

The 70/20/10 split and what actually counts as experimental

Seventy percent goes to channels that have produced opportunities at an acceptable cost for two consecutive quarters. Twenty percent goes to channels that are working but not yet proven at your target scale. Ten percent goes to things you have never run.

The rule only works if you are honest about which bucket a channel is in. A channel that produced good numbers for one quarter is in the 20, not the 70. A channel you have run for two years that has never had its cost per opportunity calculated is in neither, because you do not know what it is.

Experimental spend is where most teams cheat. A new ad format on an existing channel is not an experiment, it is optimisation. Real experiments are a new channel, a new offer type, a new audience definition, or a new motion entirely. Things that could plausibly become the next 70 percent.

Budget the 10 percent as a standing line and spend it whether or not the quarter is going well. The temptation to cancel it during a bad quarter is exactly when you most need the pipeline of future channels. Pull your creative from the SaaS ad copy templates so the experiment cost is media rather than three weeks of copywriting.

Monthly reallocation rules tied to cost per opportunity

Set the rules once, in writing, before anyone has a bad month. The point of writing them down in advance is that they get followed when the numbers are ugly and someone senior wants to move $20,000 on a hunch.

That last step is the one that separates teams who learn from teams who just move money around. Writing the prediction down before you see the outcome is the only defence against a narrative built backwards from whatever happened.

Two structural exceptions. Brand search never gets cut in a reallocation, because the cost per opportunity looks brilliant and cutting it just hands your name to a competitor. And review site contracts cannot be reallocated mid-term, so they sit outside the monthly routine entirely and get decided once a year at renewal.

Why five channels at $8K a month guarantees five failures

Because the arithmetic of the learning floor is not negotiable. At $8,000 split five ways, every channel sits below the spend level at which its optimisation works and below the volume level at which your own data means anything.

Consider what $1,600 a month buys on LinkedIn. At a $60 CPM and a 0.45 percent click through rate, that is roughly 27,000 impressions, 120 clicks, and maybe six form fills. Six. You cannot tell a good campaign from a bad one at six conversions a month, and neither can LinkedIn’s delivery algorithm, so it will not find your best pockets of audience.

The same $8,000 in one search account produces perhaps 44 leads a month at a $180 CPL, enough for automated bidding to work and enough for you to read a search terms report and act on it. One channel that functions beats five that do not, and it is not close.

The counterargument I hear is portfolio risk: what if search does not work? That risk is real and the answer is sequencing, not splitting. Run search hard for two quarters, prove or disprove it, then add the second channel with what you learned about which audiences and offers convert. The full sequencing view sits in the SaaS advertising strategy playbook.

What to cut first when the budget gets halved

It happens, usually in the week after a board meeting. Cut in this order, and cut whole channels rather than trimming every line by 50 percent.

Trimming everything by an equal percentage is the worst option available and the most common one chosen, because it feels fair. It takes a portfolio where two channels worked and turns it into a portfolio where none do. Cut whole lines, keep the survivors above their floors, and write down what you will restore first when the budget comes back.

Before the cut, run a proper audit so you know which line items are already dead. The SaaS PPC audit checklist walks the account structure, and the reporting stack that makes any of this legible is covered in PPC tools for SaaS teams.

Build your number this week

Open a spreadsheet, put your new ARR target from paid in the first cell, and work down the chain: ACV, close rate, opportunity rate, MQL rate, lead rate, CPL. The number at the bottom is your budget. If it is larger than what you have, cut the target rather than spreading the money thinner.

Then check every channel in your plan against its floor. Anything below the floor gets removed and its budget goes to the channel above it. Two funded channels beat five starved ones every quarter of every year, and the demand generation budget allocator will show you the split across the wider mix if paid is only part of your programme. If you want the broader channel mechanics behind these splits, start at SaaS PPC and paid ads.