Paid Advertising Agency

How do I know if my paid ads agency is actually doing a good job?

Most business owners running paid media through an agency are in the same position. They get a monthly report, they sit through a recurring call, the numbers in the deck look fine, and they have a quiet feeling that something is off but no real way to confirm it. They are not going to log into Google Ads and audit the account themselves, and they should not have to. The whole point of paying an agency is so they do not have to.

The good news is that you do not need to open the account to know whether the work is healthy. The signals are visible from where you already sit, and they have less to do with whether the numbers went up than with how the agency thinks, communicates, and defends its decisions.

---

The three questions that tell you almost everything

When you ask your agency a sharp question, do you get a sharp answer?

Try this on your next call. Pick the campaign with the largest spend last month and ask what specifically changed in it during the last thirty days and what the result of that change was. A real operator will answer immediately with specifics. They will tell you what hypothesis the change was testing, what the data showed, and what the next step is. An agency on autopilot will give you a general response about ongoing optimization, mention Performance Max or Smart Bidding doing its work, and steer the conversation back to the report. The quality of the answer to that one question is more revealing than any dashboard.

Does the agency talk about your business, or does it talk about the platform?

Healthy agency relationships sound like business conversations. The agency knows your close rate, your average customer value, your seasonality, your sales cycle, and your competitive landscape, and the strategy reflects that knowledge. Unhealthy relationships sound like platform conversations. Every recommendation is framed in terms of what Google or Meta wants you to do next, every change is something the platform suggested, and the agency has no clear point of view on whether any of it is right for you. The first kind of agency is operating as a partner. The second is operating as a pass-through to whatever the platform is selling that quarter.

Can the agency explain its results in the context of your market?

This is the test that catches the most underperforming agencies, because raw numbers cannot be evaluated in isolation. A small drop in lead volume against a quarter where competitors flooded the auction with new budget can be excellent work. A gain in cost per lead during an obvious seasonal lift can be ordinary work that looks good on paper. A real operator frames results against what was happening in the market, in the platform, in your category, and in your business. If the agency cannot speak fluently to those forces and tie their decisions to them, the work is being run in a vacuum, and the reporting is decorative rather than analytical.

---

What matters more than any number on the report

The conversations above will tell you a lot, but underneath them sits the work that actually determines whether the agency is doing its job. There are three layers, and most business owners never see any of them, which is exactly why most agencies get away with weak performance.

The first is what is actually being tracked as a conversion.

Ask your agency to list every conversion action set up in your account. Real conversions are sales, qualified leads, booked calls, completed checkouts. Soft conversions are page views, button clicks, time on site, scroll depth, video views, and similar engagement signals. Some agencies set up the soft ones as primary conversions and then optimize toward them, which produces excellent-looking reports and almost no business impact. Smart Bidding feeds on whatever you tell it to value. If the conversions feeding the algorithm are not the ones that matter to your business, the entire program is running toward the wrong destination. You cannot evaluate an account without first knowing what it is actually optimizing for.

The second is what happens to the numbers when branded search is removed.

This is the single most common way agencies inflate the apparent value of their work. Branded search is when someone types your company name into Google. Those people were already looking for you. They came from your existing reputation, your other marketing channels, your customer referrals, your offline presence. Capturing them with a paid ad is useful, but it is not demand creation. It is interception of demand that already existed. Agencies routinely roll branded performance into the headline numbers in their reports, because branded campaigns have low cost per click and high conversion rates, which makes the overall account averages look strong.

Ask your agency to report on non-branded performance separately from branded. The numbers will look completely different, and they will tell you whether the agency is actually driving new demand for your business or simply capturing existing demand and taking credit for it. A serious agency reports the two separately every month and welcomes the question. An agency that resists the separation is hiding something.

The third is what happens to those leads after they reach your sales team.

This is where most of the truth lives, and it sits outside the agency’s reports entirely. Volume of leads and cost per lead can both look healthy while the leads themselves are unqualified, uninterested, or unable to close. The only way to know is to track every lead from source to outcome and have your sales team rate quality honestly.

The mechanics of doing this are not complicated. A CRM with proper source attribution will tell you which campaigns produced which leads, which leads turned into opportunities, and which opportunities closed. If you do not have a CRM, we offer WordPress tracking plugins that integrate with most lead capture forms and feed the data back to you in a way you can act on. Either path is fine. What matters is that the loop is closed, the sales team is rating the leads as they come in, and the agency is reviewing that data when they make decisions about what to optimize. An agency that has never asked to see your closed-won data is an agency optimizing in the dark.

At the end of the day, the only number that matters is whether your business is growing. Lead volume that does not produce revenue is noise. Cost per lead that produces unqualified prospects is wasted spend with a friendlier label. Conversion data that ignores branded versus non-branded is a story written to make the agency look good. The agencies worth keeping are the ones that volunteer all three of these views without being asked, because growing your business is the only thing they are actually being paid to do.

---

When the right move is an outside audit

If any of this is matching what you are seeing, or if you simply cannot get clear answers to the questions above, the fastest way to know for sure is to have a senior outside operator audit the account on your behalf. An audit handled properly will tell you whether the current work is healthy, what is being missed, and what the realistic upside would be if the account were run with the depth and discipline it deserves. The cost of finding out now is almost always smaller than the cost of waiting another quarter to find out.

We do account audits for businesses that want a clear answer. The work is honest about what we find. Sometimes the existing agency is doing solid work and we say so. Often we find specific issues that have been quietly costing the business meaningful money, and the client wishes they had asked sooner.

---

What is wasted ad spend and what is it costing my business?

Wasted ad spend is the portion of your paid media budget that goes to clicks, impressions, and placements that have no real chance of producing a customer. It is the dollars flowing out of your account into searches, audiences, placements, and platforms that are never going to convert, never going to generate a qualified lead, and never going to move your business forward in any way that justifies the cost.

The published numbers on the scale of this problem are uncomfortable. The numbers we see in our own audit work are worse.

The average Google Ads account wastes between 20 and 40% of its monthly budget on clicks that will never convert, according to industry analyses. According to industry data, 73% of Google Ads accounts waste 20 to 40% of their budget. Those are the published averages.

In our own audit work, across millions of dollars of ad spend reviewed for clients ranging from mid-sized businesses to major accounts, the actual waste rates we find sit consistently between 40% and 60%. The 20 to 40% figure that gets quoted in industry articles is the floor, not the average. It is what well-managed accounts settle at. The accounts that come to us for review are coming to us precisely because they suspect something is wrong, and the audit almost always confirms that the waste is significantly higher than the public numbers suggest.

The worst case we have ever audited landed at 88% wasted spend. This was not a small business operating without expertise. This was a major account, managed by an established agency, with sophisticated reporting and a senior strategist on the call every month. Underneath the polished reporting, 88 cents of every dollar the business was spending was going nowhere a buyer existed. The owners had no idea, because the reports never made them ask the question, and the agency never volunteered the answer.

That account is not an isolated story. It is a sharper version of the same story we see across most audits. Real money, allocated against well-intentioned strategy, evaporating into searches, audiences, and placements that were never going to produce anything.

The numbers matter because the cost has two parts, and most operators only see one of them.

---

Cost number one: the dollars that left your bank account

This is the obvious one. Every click from a search that did not match your offering, every impression served to an audience with no interest in buying, every placement on a site or app where your buyers do not live, was paid for by you. The money is gone. It is sitting inside Google or Meta or LinkedIn now, and there is no path to getting it back.

The waste compounds in ways most operators do not appreciate. Fraudulent clicks and invalid traffic corrupt the historical data that Smart Bidding relies on, causing the algorithm to make systematically wrong decisions and increase bids on traffic sources that are actually fraudulent. The waste does not just cost the money it took on the way out. It teaches the platform to spend more of your budget on the same kind of waste going forward. The account learns to be inefficient, and the longer the pattern runs, the harder it is to retrain.

For most businesses, this cost is the one that gets attention because it is the easiest to see. You can pull a search terms report, look at the queries you paid for, and spot the waste. The job seekers searching for careers, the students searching for free templates, the DIY researchers, the competitor name lookups with no purchase intent, the broad match interpretations that have no relationship to what you actually sell.

But this is the smaller of the two costs.

---

Cost number two: the customers your competitors got while your budget was working the wrong angles

This is the cost most business owners never calculate, and it is almost always larger than the first one.

Every dollar of your budget that went to a wasted click was a dollar that could have gone to a buyer who was ready to purchase from your category. While your account was paying for traffic from “free legal advice” and “how to do this myself,” a competitor was buying the click from a high-intent searcher who needed exactly what your business offers. That click produced a lead. That lead became a customer. That customer produced revenue for the competitor that should have been yours.

The math compounds over time in ways most operators underestimate. If your competitor captured a $5,000 customer last quarter with a click your wasted budget should have funded, you did not just lose the $5,000. You lost the lifetime value of that customer, the referrals they would have generated, and the market share they now represent for the competitor. Multiply that across the share of your budget that is being wasted, across the quarters it has been wasted, and the cost adds up to something significantly larger than the surface number on your monthly spend.

The other dimension of this is data starvation. Smart Bidding and the algorithmic systems underneath modern paid media perform better as they accumulate conversion data. Every conversion the account misses because the budget was spent on the wrong searches is a data point the algorithm did not get to learn from. Accounts that have been run inefficiently for a year do not just have less revenue. They have a less intelligent bidding system, less refined audience modeling, and a weaker creative testing base than accounts that have been run properly for the same period. The gap between a well-managed account and a wasteful one widens over time, and closing it later requires materially more investment than running the account properly in the first place would have.

---

Where the waste comes from

The sources are predictable. The same patterns show up in nearly every audit of an underperforming account.

Broad match keywords running without disciplined negative keyword lists, producing impressions and clicks on searches that have nothing to do with what you sell. Broad match keywords without proper negative keyword coverage are the single biggest source of wasted spend in Google Ads.

Performance Max campaigns running unmanaged, distributing spend across placements that no human has reviewed and asset combinations that no human has optimized. Performance Max campaigns without account-level negatives, brand exclusions, and asset group discipline cannibalize search and waste spend on low-intent placements.

Audience targeting that is too broad, geographic targeting that includes regions where you do not serve customers, dayparting that lets the account spend at 3 AM on a Tuesday when your buyers are asleep, and device bidding that pays premium prices for mobile traffic that converts at a fraction of the desktop rate.

Conversion tracking that is misconfigured, feeding the algorithm bad data and causing it to optimize toward the wrong outcomes.

Click fraud and invalid traffic, which most accounts do not actively monitor or defend against. Industry fraud rate studies have found 65% fraud rates in photography services, 62% in pest control, and 53% in locksmith services, with broader cross-industry estimates landing in the 20 to 30% range for fraudulent or low-value clicks.

---

What an audit will tell you, with certainty

Take whatever you spend on paid media in a month. Apply the 40 to 60% waste rate that we find consistently in real audit work. That number is what is leaving your business every month with no realistic path to producing a customer. If your account is closer to the 88% end of the range we have audited before, the picture is significantly worse than that.

The only way to know which end of the range your account sits at is to have it audited by someone with the depth to find the waste and the honesty to tell you what they find. We have done this work across millions of dollars of client ad spend, and the pattern is consistent. The waste is almost always larger than the business owner expects. The opportunity to recover it is almost always larger than the cost of finding out. And the longer the account runs without that review, the more the waste compounds, both in the dollars lost and in the customers the competitors are quietly taking instead.

---

How much should a business spend on paid advertising?

This is the question every business owner asks first, and there is no single number that answers it honestly. The right budget is not a percentage of revenue or a benchmark pulled from an industry report. It is the output of a calculation that ties what you are willing to spend to what you are trying to produce, and that calculation looks different for every business depending on goals, risk tolerance, and the math underneath the sales process.

What follows is how we actually think about this with clients, because the textbook answer of “spend 10% of revenue on marketing” is the kind of advice that wastes money on businesses that should be spending more, and overcommits businesses that should be spending less.

---

The conversation starts with what you are trying to do

There are three different versions of this conversation, and the right budget for each one is materially different.

Some clients want to keep the business busy at a steady state. They have a healthy operation, predictable demand, and they want paid media to maintain a consistent flow of inquiries without dramatic month-to-month swings. The budget conversation here is about steady-state pacing, defensive coverage of the searches that matter most, and protecting market share against competitors who are spending more aggressively. This is the most common version, and it is also the version where most agencies quietly underdeliver because the work is less visible than a scaling campaign.

Some clients want to scale. They have capacity to handle more, sales infrastructure to convert what comes in, and they want paid media to push hard against the ceiling. The budget here is about how fast they want to grow and how much risk they are willing to absorb to get there. Scaling spend produces lower marginal returns than steady-state spend, because the highest-intent traffic is captured first and the additional volume comes from progressively less qualified audiences. That is not a flaw in the strategy. It is the cost of expansion, and clients who understand this in advance make better decisions about how aggressively to push.

Some clients want to operate within a fixed budget. They have a number that finance has approved, and the conversation is about how to extract the maximum result from that specific number. This is the most disciplined version, and it requires the agency to make hard prioritization decisions about which campaigns get funded, which channels get tested, and which opportunities have to be set aside for the next quarter.

The honest answer to “how much should I spend” depends on which of these three you are. The same business at different points in its lifecycle can sit in different categories, and the budget should reflect that.

---

The math that turns goals into a real budget number

Once you know what you are trying to do, the budget calculation becomes concrete. For most businesses with a sales process, it works like this:

Start with a target. How many new customers do you want to acquire next quarter, or next year? This is the number paid media is being asked to contribute toward. If your business needs 40 new customers a quarter and paid media is one of four contributing channels, the share you are asking paid to produce is 10 customers a quarter.

Work backwards through your close rate. If your sales team closes 25% of qualified leads, producing 10 customers requires 40 qualified leads. If your close rate is 40%, you need 25. If it is 10%, you need 100. The close rate has more leverage on the budget conversation than almost any other variable, and most business owners do not have it accurately tracked.

Apply the cost per lead for your category. This is where benchmarks become useful. WordStream’s 2026 data shows the average cost per lead in Google Ads is $66.69, with the spread ranging from $26.84 in arts and entertainment to $131.63 in legal services. Your category has its own number, and once you have it, the math is simple. 40 qualified leads at $80 cost per lead is $3,200 in spend per month, or roughly $10,000 a quarter, to produce the lead volume that will deliver 10 customers at your close rate.

Compare against customer value. If your average customer is worth $5,000 over their lifetime, spending $1,000 in ad cost to acquire one is excellent. If they are worth $500, the same $1,000 is a loss. The acceptable cost per acquisition is a function of customer lifetime value, and businesses that have not calculated theirs are flying blind on the entire spend question.

The number that falls out of this math is the budget that matches your goals, your sales infrastructure, and the economics of your business. It is not a guess. It is the answer to a specific question, and the question is more important than the answer.

---

Where risk tolerance changes the picture

The other variable that shapes the budget conversation is how much risk a business owner is willing to absorb in pursuit of growth.

Some clients want paid media run conservatively. They want predictable performance, tight control over spend, and the security of knowing the number is not going to surprise them. The budget reflects that. The work happens within disciplined guardrails, growth is incremental, and the downside risk is small.

Some clients want paid media run aggressively. They are willing to accept lower marginal returns on the next dollar of spend if it produces meaningfully more volume at the top. They understand that scaling requires accepting a higher cost per acquisition in exchange for capturing more of the market. The budget for these clients is larger, the testing is more ambitious, and the conversations are about how much faster the business can grow rather than how to keep things steady.

Neither approach is correct in isolation. The right approach is the one that matches the owner’s appetite, the business’s stage, and the operational reality underneath. The wrong approach is committing to scale spend when the sales team cannot handle the volume, or committing to conservative spend when the market opportunity rewards aggression.

---

Why most budget conversations go wrong

The most common failure is that the budget gets set before any of this math gets done. Finance approves a number based on what was spent last year, the agency commits to producing results against it, and nobody steps back to ask whether the number matches the actual goals of the business.

The second failure is that the budget never gets revisited as the math changes. Cost per lead drifts. Close rates shift. Customer value evolves. A budget that made sense eighteen months ago can be significantly under or overcommitted today, and the only way to know is to run the math again. Most accounts have not had this conversation in years.

The third failure is that the budget is treated as fixed instead of as a function of opportunity. If the math shows that the next $5,000 of monthly spend would produce $30,000 of incremental revenue at acceptable margins, the right answer is to find the $5,000, not to declare the budget closed. If the math shows the next $5,000 would produce $4,000 of revenue, the right answer is to redirect it elsewhere. The budget should respond to what the data is showing, not sit as a fixed line item that the agency works against regardless of the underlying economics.

---

How we actually approach this with clients

The conversation we have with new clients is not about how much they want to spend. It is about what they want the business to look like in twelve months and what role paid media can realistically play in getting them there. From there, we work the math backwards through close rate, cost per lead, customer value, and the operational capacity the business has to handle the resulting volume. The number that comes out of that exercise is the budget we recommend, and it is almost always different from the number the client walked in with.

Sometimes the recommendation is to spend more, because the math shows the business is underinvested relative to the opportunity and is leaving growth on the table. Sometimes the recommendation is to spend less, because the math shows the current budget is producing diminishing returns and the same outcome could be achieved more efficiently. Sometimes the recommendation is to hold the budget steady and change the allocation, because the issue is not how much is being spent but where the spend is going.

Whatever the recommendation is, it comes from the math, not from a default percentage or a benchmark pulled from a blog post. That is what separates a real budget conversation from the kind that produces the average waste rates the industry is built on.

---

How is AI changing paid advertising, and what does it mean for businesses running ads today?

AI is now embedded in every major paid advertising platform, and the change is not incremental. Smart Bidding decides how much to pay for every click. Performance Max chooses which channels to spend on, which audiences to target, and which creative combinations to serve. Meta’s Advantage+ runs the same kind of autonomous loop across Facebook and Instagram. AI Max for Search expands your keywords and dynamically rewrites your ad copy on the fly. Google’s AI now controls more of the tactical execution layer than it ever has before, and Smart Bidding strategies like Target ROAS and Target CPA are no longer optional advanced tactics, they are the baseline expectation from Google’s own systems.

The platforms are not subtle about the direction this is going. By early 2026, Meta has begun defaulting new campaign creation toward Advantage+ settings, with targeting, placements, creative enhancements, and budget optimization all pre-selected. 65% of Meta advertisers are already scaling through Advantage+. 73% of Google Ads accounts now use automated bidding strategies, compared to 31% in 2022. The era of manual bidding and tightly controlled targeting is over for most accounts, and the era of governing AI systems has begun.

The question for business owners is no longer whether to use AI in paid media. The question is how to govern it well, when to lean into it, and when to push back against the platform’s defaults. That is where most of the value is being created right now, and it is also where most of the damage is being done.

---

When AI is a genuine multiplier

When AI in paid media is working properly, the results are real and the math is favorable. Performance Max campaigns generate 35% more conversions at 20% lower CPA compared to equivalent manual campaigns, according to Google’s Q1 2026 data. Advantage+ Shopping campaigns deliver an average ROAS of 4.52x compared to 3.70x for manually managed campaigns, a 22% improvement. Advertisers using Smart Bidding see 20 to 40% better performance than manual bidding, according to Google’s internal data.

These numbers are not propaganda. They are achievable when the underlying conditions are right. AI in modern paid media is processing thousands of signals per auction in milliseconds, including device, location, time of day, search context, browsing history, and audience membership. No human can compete with that on volume of decisions, and no senior operator should try to. The right posture is to let the AI handle the tactical execution it is designed for, and to focus human expertise on the inputs and the strategy that determine whether the AI produces good outcomes or bad ones.

The catch is in the word “when.” The platform-reported uplifts assume the AI is being fed clean conversion data, governed by experienced operators, and pointed at the right objectives. None of that is the default state of most accounts.

---

When AI is dangerous

The same systems that produce the strong results above will quietly destroy a budget when the inputs are wrong. AI does not have judgment. It has signals, and it optimizes toward whatever signal it is told to value. If the signals are bad, the AI executes the wrong strategy with extraordinary efficiency.

The most common failure mode is conversion data quality. Modern ad platforms now rely more on automation than on manual targeting, which means the quality of your data signals is critical, and Meta and Google’s AI systems need high-quality, server-side and CRM-backed signals to perform well. Accounts that count low-value events as conversions, miss conversions because tracking is broken, or feed the AI inconsistent first-party data, are training the algorithm to optimize for the wrong outcomes. When signal quality declines, optimization becomes guesswork, and return on ad spend inevitably suffers. The account looks like it is being managed, because the AI is running. It is just running toward the wrong destination.

The second failure mode is lack of governance. AI systems left unmanaged drift toward whatever produces the easiest conversions, which usually means spending more on existing customers and brand search than on new customer acquisition. Without an existing customer budget cap, Advantage+ will naturally spend heavily on existing customers because they convert at higher rates, which is why setting the cap at 20 to 30 percent forces the algorithm to find new customers. The same dynamic applies inside Performance Max, which will cannibalize search campaigns and feed on branded queries if the proper exclusions are not in place. Performance Max campaigns without account-level negatives, brand exclusions, and asset group discipline cannibalize search and waste spend on low-intent placements.

The third failure mode is creative starvation. AI in modern paid media is creative-led. The platforms test combinations, learn what resonates, and optimize toward the assets that perform. Accounts with thin creative libraries are giving the AI nothing to work with, and the algorithm produces correspondingly thin results. Creative is now the primary performance lever in paid media, and one dental practice that created 12 different creative variations saw their cost per new patient drop from $320 to $95 on the same budget and targeting. Without diverse creative inputs, the AI is being asked to optimize a system with no levers.

The fourth failure mode is platform concentration risk. As Meta Advantage+ becomes more capable, the temptation is to consolidate more and more ad spend into a single automated system, which creates meaningful dependency on a single platform, and a prudent approach maintains diversified paid media across Meta, Google, and emerging channels. The same applies to Performance Max within Google. Accounts that have surrendered all execution to one or two AI systems have no insurance against a platform policy change, algorithm shift, or pricing change that could disrupt their pipeline overnight.

---

What separates operators who get AI right from operators who do not

The difference between AI working for you and AI working against you comes down to four things, and none of them are tactical settings inside the platform.

Conversion data integrity. The right operators audit conversion tracking obsessively. They install Conversions API for Meta, Enhanced Conversions for Google, and offline conversion imports to feed the platforms the actual revenue from closed deals. They configure the right events as primary conversions, they monitor data freshness, and they fix tracking problems before the AI has time to learn from bad data. Meta recommends running both Pixel and Conversions API simultaneously with matching event names and Event IDs for deduplication, and the operators who do this consistently see materially better results than those who do not.

Governance, not surrender. The right operators treat AI as an extremely capable junior staffer that needs supervision. They set the goals, they define the guardrails, they monitor for drift, and they intervene when the AI starts spending in places it should not. They do not let Performance Max run unmanaged. They do not assume Advantage+ will figure it out. They are watching the outputs constantly and pulling the levers that exist to keep the system aligned with the business objective.

Creative as a discipline. The right operators treat creative development as the highest-leverage activity on the account. They produce diverse assets, test relentlessly, and feed the AI a steady stream of new inputs. They understand that in 2026, the creative is doing more of the performance work than the targeting, and the budget for creative production reflects that.

Strategic judgment about when AI is the wrong choice. Sometimes the right answer is to opt out of the AI feature the platform is pushing. Sometimes Manual CPC is the right bid strategy for a campaign with insufficient conversion volume. Sometimes Standard Search is the right campaign type instead of Performance Max. Sometimes the platform’s recommendation is wrong for your business, and an experienced operator recognizes it and refuses. The default of accepting every AI recommendation is the path most accounts take, and it is the path that produces the average waste rates the industry is built on.

---

What this means for businesses running ads today

The implication is straightforward. AI has made paid advertising both more powerful and more dangerous than it has ever been. The accounts that are run by senior operators who understand how to govern AI systems are producing the best results the industry has ever seen. The accounts that are run on autopilot, with the AI features turned on and the human supervision turned off, are bleeding budget faster than they ever have.

The business owner’s question is no longer whether to embrace AI in paid media. The question is whether the people running your account have the depth to use it properly. AI has not replaced the need for senior expertise. It has raised the stakes on it. The gap between an account run by someone who understands what the AI needs to succeed and an account run by someone who hopes the AI will figure it out, has never been larger. That gap is where the next several years of competitive advantage in paid media will be decided.