Why Most SMBs Are Burning Their Marketing Budget on the Wrong Channels

A small-business owner we worked with recently was spending about $11,000 a month on paid media. Three channels, one agency, two tidy dashboards. Every quarterly review showed the same thing: a healthy return on ad spend, a pipeline that looked full, and one chart in particular that everyone liked because the line went up and to the right. The numbers were not made up. The numbers were just measuring the wrong thing. When we ran a four-week GEO holdout test on the highest-spend channel, revenue in the paused region barely moved. The dashboard had been crediting that channel for revenue it was not actually producing. The owner had been paying $11,000 a month to confirm something that would have happened anyway.

That story is not unusual. It is the median story. DemandScience's 2026 State of Performance Marketing Report surveyed 750 senior marketing leaders at companies between $100M and $5B+ and found that the average team wastes about 25% of its marketing budget on activities that produced no measurable results. Teams whose dashboards are most often misleading waste 30%; teams with reliable measurement waste 23%. The dashboards are not innocent bystanders. They are correlated with the very waste they are supposed to prevent.

This paper is about where SMB marketing money actually leaks, why nobody in the value chain has a strong incentive to point it out, and what an owner or marketing director can do this week to find their own leaks. The answer is not "spend more." For most SMBs, the answer is "spend the same amount, more honestly."

Why is the SMB marketing budget shrinking when companies need it most?

Marketing budgets across the broader market are flat or compressing. Gartner's 2025 CMO Spend Survey found marketing spend has flatlined at 7.7% of overall company revenue, with 59% of CMOs reporting they do not have enough budget to execute their strategy. Mid-market and SMB teams typically run on a tighter ratio than that, not a more generous one.

Inside that flat envelope, the mix is tilting. Paid media's share of CMO budget rose to 31% in 2025, up from around 28% in 2024. Martech, agencies, and labor all declined to make room. That single shift, more paid media inside a flat budget, is the headline story of SMB marketing in 2026. Less is going to the people producing creative, the platforms storing customer data, and the agencies running campaigns. More is going directly to ad networks.

This matters for waste arithmetic. When a fixed share of paid media is incremental and the rest is captured demand, growing the paid-media slice grows the wasted slice in lockstep. A team that wasted 25% of a 28% paid budget is now wasting 25% of a 31% paid budget. The percentage is the same. The dollar amount is bigger. Most SMBs do not run the math.

Add to that the labor squeeze. The Fall 2024 CMO Survey from Duke Fuqua's Christine Moorman put marketing budgets at 7.7% of revenue, with B2B product companies at 6.4%, B2B services at 9.0%, and B2C product companies at 15.5%. Marketing teams are not getting bigger. The work expected of them is. HubSpot's 2026 State of Marketing report found 83.5% of marketers say they are expected to produce more content than they the year before, and 35.7% say "much more." A flat budget plus a heavier content load plus a paid-media tilt is a recipe for hurried decisions about where money goes. Hurried decisions are where waste lives.

Where does the money actually leak?

Across the SMB and mid-market teams we work with, four leaks recur often enough to be near-universal.

Leak one: branded search bidding the brand was already going to win. A team runs Google Ads on its own brand name. The dashboard reports a 15x return. The team scales the budget. The reality is that most of those clicks were searches by people who already knew the brand and were going to convert through organic results. Bidding on branded search has a defensible role, mostly to crowd out competitors, but at most SMBs the spend has crept well past that purpose. The honest test is to pause branded search bidding for two weeks and watch organic branded clicks. They almost always rise to absorb most of the lost paid clicks. The incremental contribution of branded search is usually a fraction of what the dashboard shows.

Leak two: retargeting display at scale. Display retargeting works in moderation. It stops working when the volume gets large enough that the same buyers see the same ad twenty or thirty times. The dashboard still credits the display network for the eventual conversion (last click or position-based attribution), so spend keeps rising. The buyer would have come back anyway. Most SMB display budgets could be cut 40% with no measurable revenue impact, and the cut would surface fast in any incrementality test.

Leak three: paid social campaigns that capture demand instead of creating it. Meta and TikTok do an excellent job of finding people likely to convert. The problem is "likely to convert" mostly means "already in-market." A team running Advantage+ Shopping campaigns may be paying Meta's algorithm to deliver buyers who would have arrived through organic channels. Meta credits itself, the dashboard agrees, the spend grows. The cleanest way to spot this leak is geo holdout testing. We have seen first holdout tests show 35% to 60% less incremental revenue than the in-platform reporting claimed. The number is uncomfortable. It is also more honest than anything the platform was telling you.

Leak four: agency retainers without defined output. Many SMBs pay $5,000 to $15,000 a month for an agency relationship without a clean monthly deliverable list. The retainer becomes a fixed cost; the work fluctuates; the team never compares the two. Gartner's 2024 CMO Spend Survey found 39% of CMOs plan to cut agency budgets, with the top action being "eliminate unproductive agency relationships." Most agencies are not unproductive. They are unspecified. The fix is a written deliverable list, monthly, with named owners and dates. For the deeper math on this trade-off, see our agency vs. AI marketing cost comparison.

These four leaks together account for most of the 25% waste figure. The good news: each one is fixable in a single quarter without firing anyone or changing platforms.

Why do dashboards hide these leaks?

Three structural reasons.

The first is that in-platform dashboards are written by the platforms whose performance they grade. Meta's reporting credits Meta. Google's reporting credits Google. TikTok's reporting credits TikTok. If you sum the revenue every platform claims it generated for an SMB, the total often exceeds the company's actual revenue. We have seen audited stacks where the implied multiplier across reporting was close to two. Each platform is technically not lying. Each platform is using attribution rules favorable to itself. The arithmetic does not add up because the platforms were never required to add up to one.

The second is that multi-touch attribution models, the ones built into most marketing dashboards, were designed in a different world. They assumed they could see most of a customer journey through cookies and device IDs. They cannot anymore. Apple's App Tracking Transparency made user-level mobile tracking opt-in, and three years later AppsFlyer's data shows the global opt-in rate has stabilized at around 50% (44% in the US). About 84% of gaming app developers and 68% of non-gaming developers now show the prompt. Half of the customer journey, on mobile, is invisible to the systems that used to measure it. The third-party cookie story added another layer of confusion: Google's Chrome team reversed the planned cookie deprecation in July 2024, abandoning the user-choice prompt entirely, while Safari and Firefox already restrict cross-site tracking by default. The dashboards still produce confident reports anyway, because reporting "we cannot fully measure this" is not what platforms are paid to say.

The third is dashboard sociology. Marketing leaders take dashboards into board meetings. A clean waterfall chart is easier to defend than "we cannot fully measure our channel mix." So the chart wins, even when the chart is fiction. Owners and CMOs both know this on some level and keep using the chart because the alternative is harder to talk about.

None of this is malicious. It is a system where the people producing the numbers, the people consuming the numbers, and the people selling the platforms behind the numbers all have soft incentives to keep using numbers that are not quite right. The first step toward a less wasteful budget is being willing to say so out loud.

How big is the leak at a typical SMB?

A useful exercise: run the math on your own situation against the DemandScience 25% baseline.

A SMB spending $20,000 a month on marketing is wasting roughly $5,000 a month, or $60,000 a year, on activities that produce no measurable revenue. That figure could fund a full-time content marketer, a full retargeting platform, six months of a high-end SEO program, or a year of CRM and email infrastructure. The waste is not theoretical. It is one full-time hire's salary, every year, getting set on fire.

The teams hit hardest by misleading metrics waste 30%, which on the same $20,000 monthly spend is closer to $72,000 a year. The teams with reliable measurement waste 23%, or about $55,000. The gap between best and worst is roughly $17,000 a year on a $240,000 annual spend. That gap is real money for an SMB. It is also entirely controllable through measurement hygiene, not budget changes.

The math gets worse when you multiply across channels. Most SMBs run five or six paid channels in some form. Each channel has its own dashboard, its own attribution model, and its own incentive to credit itself. Each channel is allowed to over-claim a few percent and still look correct. The compound over-claim across the stack is how the numbers stop adding up.

What does an honest marketing budget allocation look like for an SMB?

There is no universal correct mix. The right allocation depends on business model, sales motion, average customer lifetime value, and how mature the brand already is. But a useful rule of thumb for SMB and mid-market teams: paid media should be a meaningful slice, not the whole pie.

A balanced 2026 SMB marketing budget that we see working across our customer base typically looks something like this:

• 30% to 40% paid media (down from the industry-wide 31% only because most SMBs are running closer to 50% to 60% paid)

• 20% to 25% content production (organic search, blog, video, social, email content)

• 15% to 20% owned-channel infrastructure (CRM, email platform, website, analytics)

• 10% to 15% creative and brand work (positioning, design, brand voice systems)

• 5% to 10% experimentation budget (incrementality tests, new channel pilots, audience research)

The ratio matters less than the principle: every dollar that goes to paid media buys a single month of attention. Every dollar that goes to content, owned infrastructure, and brand work buys attention that compounds for years. Most SMBs are spending too much in the first column and starving the second three. The dashboards reward the first column because the first column is the most measurable. That measurability bias is the silent shape of most SMB marketing waste.

The shift does not have to happen overnight. A team that takes 5% of paid budget and moves it into content production for two consecutive quarters will, by the end of year, have meaningfully more organic search volume, more direct traffic, and more dark-social referrals. The new traffic is not as easy to credit cleanly. It is also not as easy to lose.

What can an owner or marketing director do this week?

Five moves, in priority order, that we have seen work for SMBs across industries.

1. Add a "How did you hear about us?" field at every conversion point. Forms, post-purchase surveys, sales-call intake notes. Make it required. Aggregate the answers monthly. Within 60 days you will have a parallel data source that does not depend on cookies, ATT, or in-platform reporting. Most SMBs running this for the first time discover that 30% to 50% of new customers come from channels their dashboards do not track at all. The discovery itself is worth the small UX cost of one more form field.

2. Run a four-week geo holdout test on your highest-spend paid channel. Pick two comparable cities or regions. Pause spend in one. Compare revenue. The number you get is the truth. The number on your dashboard is a marketing claim. If the truth is meaningfully lower than the claim, scale the channel down until the two numbers agree.

3. Write a one-page list of every monthly deliverable your agency owes you. Specific deliverables, named owners, dates, and acceptance criteria. If the agency cannot or will not produce that list, that is the answer. If the list exists and is not being met, you have grounds for the renegotiation conversation. Most SMB-agency relationships drift because nobody wrote the list.

4. Audit your branded-search spend. Pull the last 90 days of branded-keyword spend. Estimate, even directionally, what share of those clicks would have come through organic results without bidding. For most SMBs, the answer is 60% to 80%. Cut branded-search spend by half for a month and watch what happens. If revenue holds, leave it cut.

5. Cap retargeting frequency. Most SMB retargeting campaigns are running with no frequency cap or a very loose one. The same buyer is seeing the same display ad dozens of times. Set a frequency cap of three to five impressions per buyer per week. Most teams see no revenue impact and a 30% to 50% reduction in display spend. The savings move to channels that compound.

None of this requires new software. None of this requires firing anyone. None of this requires a transformation project. It requires a willingness to look at the numbers without flinching.

Where does AI fit in?

AI is on every marketing-vendor pitch deck right now. The honest answer is that AI helps with some pieces of the SMB marketing problem and does not help with others.

Useful: producing the volume of content the modern channel mix demands. The 83.5% of marketers under content-volume pressure cannot get there manually inside an SMB headcount. AI-assisted content production, when paired with a real brand voice system, gets a small team to enterprise output without enterprise hiring. AI is also good at categorizing self-reported attribution responses, normalizing UTM data, drafting first-pass campaign briefs, and surfacing patterns in CRM data the team would never have time to read.

Not useful: replacing the underlying measurement. AI cannot invent the incremental signal that ATT and walled gardens took away. Anything that promises to use AI to reconstruct the customer journey from sparse data is doing the same wrong math the old multi-touch platforms did, just with a more confident interface. A confident wrong answer is more expensive than an honest uncertain one.

The teams winning right now are not the ones with the smartest measurement. They are the ones producing more on-brand, channel-native content per week, feeding their own first-party data flywheel, and getting cited in the conversations measurement cannot see. Better content makes the budget conversation easier because the wins are bigger and the leaks are easier to spot.

A short, honest soft sell

FUEL is a marketing platform for SMB and mid-market teams that need to produce more content, in their own voices, across more channels, without growing headcount or burning the budget on tools that overlap. We are not an attribution platform. We do not promise to reconstruct the customer journey. We do help two and three-person marketing teams produce 30 days of content in an afternoon, in a brand voice that sounds like the company instead of like an AI tool.

If you read this paper and recognized your own dashboard, the most useful next step is probably not to buy another measurement platform. It is to make sure the content engine feeding your channels is producing enough volume, in your voice, that the budget you are spending matters in the first place. The waste problem is real, and most of it traces back to channels the dashboards over-credit. The opposite problem (good channels you cannot scale because you cannot produce enough content for them) is the one most SMBs do not realize they have.

If you are a business owner, run the Foundation Report on your business. If the output surprises you, that is the point.

If you're an agency, generate a Foundation Report on a client you have worked with for years. If the output does not challenge your thinking, walk away. If it does, the team plans are priced for agencies ready to scale what works.

Generate My Foundation Report

If a different paper in the series matches where you are right now, the full list is at /white-papers.