Paid media is the only channel where a pricing model changes the advice you get. A B2B PPC agency paid a percentage of your spend has a structural reason to prefer bigger budgets. One paid a flat fee has a structural reason to prefer efficiency and a structural reason to resist scope growth. Neither is dishonest. You just need to know which pressure you are buying, and what the number covers.
This guide breaks down the three real fee models, runs the total cost math at three spend levels, finds the point where an in-house buyer wins, and lists the terms that decide what you keep when the engagement ends. The channel-level context, including realistic spend floors, sits on our B2B advertising agency page.
The three fee models, and what each one rewards
Almost every proposal you will read is one of three structures wearing different language. Identify the structure first, then compare numbers. If you are about to send a formal request out, our free B2B paid advertising RFP template contains the exact fee and ownership questions that force these structures into the open.
Percent of media spend. Typically 10 to 20 percent of monthly media, usually with a minimum fee so small accounts remain viable. It scales with your budget, which is defensible because pacing and structure work does grow with spend. The incentive problem is specific and predictable: the month your data says to cut a channel in half is the month the agency's revenue falls. Good firms have that conversation anyway. You should still know that you are asking them to.
Flat retainer. A fixed monthly fee, commonly $3K to $15K depending on channel count and how much of the surrounding work is included. The fee is independent of spend, so recommending less spend costs the agency nothing. The failure mode is the mirror image: if your budget triples, the work grows and the fee does not, and service quality quietly degrades unless the contract has a review trigger.
Hybrid. A smaller base retainer, often $3K to $8K, plus 5 to 10 percent of media above a defined threshold. This is the most balanced structure for programs expected to scale materially inside the term. It adds one more number to negotiate, which is a fair trade.
A fourth model, performance or per-lead pricing, deserves a warning rather than a row. It rewards volume of whatever event is counted, so it optimizes toward cheap form fills, and every dispute becomes an argument about whether a lead qualified. For considered purchases with committees and long cycles, it usually costs more in wasted sales time than it saves in fees.
Total cost of ownership at three spend levels
Fees are only comparable when you add everything the fee excludes. The math below uses the middle of the observed ranges and assumes a two-channel program with landing pages and measurement genuinely covered. Treat it as a template to run with your own quotes, not as a benchmark.
At $15,000 monthly media. Percent of spend at 15 percent is $2,250, but most firms apply a $4,000 to $5,000 minimum at this level, so you effectively pay a flat fee with a percentage label. A flat retainer of $5,000 is directly comparable. Add roughly $500 per month in tooling for call tracking, landing pages, and reporting. Total cost of ownership lands near $20,500, meaning about 27 percent of your outlay is going to the people and tools rather than the auction. That ratio is the number to watch.
At $40,000 monthly media. Percent of spend at 15 percent is $6,000. A flat retainer for the same scope is often $8,000 to $10,000, because more channels and more creative volume are in play. Here percent of spend is usually the cheaper structure on paper, and the question becomes whether you trust it when the data says to cut. Total cost of ownership runs $46,500 to $50,500, with the people-and-tools share down to roughly 15 to 20 percent. This is the level where the model choice starts to cost real money.
At $100,000 monthly media. Percent of spend at 12 to 15 percent is $12,000 to $15,000 per month, or $144,000 to $180,000 a year in fees. A senior in-house paid media manager plus employer costs and tooling generally lands under that, which is why programs at this level tend to internalize buying and retain outside help for strategy, creative, and measurement. The honest caveat is coverage: one hire cannot replace a team across search, LinkedIn, creative, and analytics, so the real comparison is one hire plus contractors against one retainer.
- Watch the ratio of fee plus tooling to total outlay, not the fee in isolation
- Under roughly $25K monthly media, percent of spend and flat fee usually converge because of minimums
- Between $25K and $75K, percent of spend often looks cheaper and carries the incentive cost
- Above roughly $75K, run the in-house comparison seriously
The break-even point against an in-house buyer
The in-house case is strongest when spend is sustained, channels are few, and the work is steady-state optimization rather than repeated new builds. Under those conditions a single senior buyer at a loaded cost of roughly $130K to $180K a year, plus $6K to $15K in annual tooling, beats a 12 to 15 percent fee somewhere around $75K to $100K in monthly media.
Three costs get left out of that case with predictable regularity. Hiring and ramp take three to six months, and a funnel does not pause while you interview. One person is a single point of failure, so the program stops during vacations and stops entirely at resignation. And a buyer is not a creative team or an analytics engineer, so you either add contractors or accept that testing velocity drops. The durable pattern we see most often is one internal owner holding strategy and budget with an outside team running execution, which is the same conclusion the firm versus in-house analysis reaches for marketing more broadly.
What the fee should include, and the three usual exclusions
Two proposals at the same monthly number can differ by an entire function. Get the inclusion list in writing before comparing prices at all.
- Landing pages. Often excluded, and often the actual constraint. An agency optimizing traffic into a page it cannot edit is bidding against your own conversion rate.
- Creative production. Ad copy is usually included. Design volume, video, and brand campaigns usually are not, and B2B social burns creative faster than teams expect.
- Measurement plumbing. The form handoff, CRM fields, offline conversion imports, and the written definition of a qualified conversation. Skip this and the platforms optimize toward form fillers, which is exactly what you will get.
Everything else should be inside the fee: ICP-based audience definition, account structure, offer and ad testing, bid and budget pacing, negative and exclusion hygiene, and monthly reporting with a named person who wrote it. Our own approach bundles paid with the site and content work it depends on, described on the paid advertising service page, and the channel-specific playbook for software companies is in the B2B SaaS paid advertising guide.
Red flags in a paid media proposal
These are the patterns that most reliably predict a bad engagement, in rough order of severity.
- The agency's business manager owns the ad accounts, pixels, or audience lists
- Media spend runs through the agency's card and appears with a markup or an unexplained blended figure
- Fee is quoted only after a qualification call, with no published range anywhere
- No named buyer, or a pitch team that will not be doing the work
- Sample reporting that stops at impressions, CTR, or cost per lead
- Guaranteed lead volume or guaranteed cost per lead before any account access
- Percentage case studies with no baseline, spend level, or time window
- No written out-of-scope list, and reluctance to produce one
Any one of these is a question. Any two together is usually a reason to move on, the same threshold we apply in the agency evaluation scorecard.
Common mistakes buyers make on paid pricing
Most bad paid outcomes trace back to a purchasing decision rather than a bidding decision.
- Optimizing the fee down until the media is starved. A $2K fee on a $6K budget spread over four channels produces four inconclusive tests. Fewer channels with more spend beats a cheaper manager.
- Comparing a flat fee to a percentage at today's spend only. Model both at your planned spend in month nine, not month one.
- Treating the ad account as the agency's asset. Conversion history and audience lists are months of compounding value that should not leave with the vendor.
- Buying paid to fix a positioning problem. Paid buys faster confirmation that the offer is unclear. It does not clarify it.
- Scaling spend on a promising week. Signal needs roughly 30 conversions per audience and offer. Scaling earlier just funds the wrong structure faster.
- Leaving the qualified conversation undefined. Without it, month three becomes a debate about lead quality instead of a decision about budget.
A rubric for comparing two paid proposals
Score each finalist from 1 to 5 on the six dimensions below, then multiply by the weight. The exercise takes an afternoon and produces a defensible decision.
- Total cost of ownership at planned spend, weight 3. Fee plus media plus tooling plus whatever the fee excludes, modeled at month nine.
- Incentive alignment, weight 3. Does the fee move when the right call is to spend less?
- Asset ownership, weight 3. Accounts, pixels, audiences, creative files, and pages in your name, with a written handover.
- Reporting endpoint, weight 2. Cost per qualified conversation and pipeline influenced, with the definition agreed pre-launch.
- Named buyer and account load, weight 2. Who runs it, and how many accounts they carry.
- Written exclusions, weight 1. A specific out-of-scope list, produced without prompting.
If two finalists tie, choose the one whose exclusions cost you least to absorb internally. That is the same tie-breaker that decides most agency selections, as we argue in what actually makes an agency the best fit.
How we price paid, and when we are the wrong answer
We charge a flat retainer and never a percentage of spend. Your media is billed to you directly with no markup, your ad accounts and pixels stay in your name, and paid runs inside an integrated program alongside the website, content, and search work it depends on, because in software the paid number is usually capped by the page and the offer rather than the auction. Our tiers are published on the pricing page, all on a 6-month minimum.
We are the wrong answer if you want paid executed in isolation, in which case a dedicated paid shop will be cheaper for the same quality. We are also the wrong answer if your available media budget is under roughly $5K per month and your deal size is small, because that money will do more work in organic and product-led motion first. And if what you need is cold outbound conversations this quarter, that is a different motion and we refer software companies to a specialist outbound partner.
If you want an outside read on the proposals in front of you, bring the two quotes and your current reporting to a 30-minute growth call. We will model the total cost at your planned spend and name the model that fits, even when the answer is to keep the partner you have.