How Much to Spend on Marketing: A Number You Can Defend

You have three proposals on the desk. One is a flat monthly figure that covers "everything". One separates a management fee from an ad budget you would fund yourself. One is cheaper than both and lists a set of deliverables you cannot quite picture. Somebody has told you the answer is a percentage of revenue, and when you worked it out the three quotes landed above, below and roughly on it, which settled nothing. The number you are looking for is one you can compare against the others today and against your own records six months from now, and building that is a different job from picking a percentage.

Three figures circulate as the answer, and each is a real number from a real survey with a population that is visible once you look. BDC, the Canadian development bank, published a survey in 2012 finding that marketing and sales budgets tended to sit between 2% and 10% of revenue, higher for firms with a few million in sales. It combined sales with marketing, and 30% of respondents spent under $5,000 a year. Gartner's 2025 CMO Spend Survey reported 7.7% of revenue, from 402 marketing leaders in North America, the UK and Europe, the great majority at organisations above $1 billion in revenue. Intuit's 2025 small-business advertising report estimates an average advertising budget of $78,000, from 1,006 owners and marketing leaders, with the advertising question excluding labour and advertising averaging 37% of respondents' total marketing budgets.

A Canadian combined figure from fourteen years ago, a large-company figure, and an advertising-only figure with labour stripped out cannot be averaged into a rule for a plumber in Arlington. What they can do is give you a sense of scale to sanity-check a proposal against, and that is the most any benchmark does. The number that describes your business is the one you build from your own revenue, your gross margin per job and what a customer is worth to you over time.

The reason three quotes cannot be ranked is that they bundle different things. Intuit's own survey separates advertising from agency fees, website production and software subscriptions, and a proposal that folds all four into one line is not cheaper or dearer than one that itemises them. It is just unreadable until you unpack it.

• The recurring service fee: what the provider charges for their time each month, and for what scope.
• Media spend: the money that goes to Google, Meta or anyone else for placement. Ask whether it is inside the monthly figure or on top, and who pays the platform directly.
• Software: email tools, scheduling, reporting, tracking. Often passed through at cost, sometimes marked up, occasionally absorbed.
• One-time work: the website changes, the tracking setup, the creative production, the strategy document. Front-loaded, and it should be priced as front-loaded.
• Your own hours: approvals, supplying facts and photographs, answering the leads. Intuit's respondents excluded labour from their advertising budgets, and you can make the same choice, but make it consistently across all three quotes.

Once each proposal is written out in those five lines, you will usually find the "everything" quote is mostly media spend, the itemised one is mostly fee, and the cheap one is cheap because a line is missing. None of that counts against any of them. It is the first moment the three can be looked at together.

The costliest mistake in this decision is not overpaying. It is committing to a monthly number with no record of what the business looked like before it started, and no agreed definition of what a good month is. Six months later there is a report full of clicks and a feeling, and the renewal decision gets made on the feeling.

Google Ads calls the thing you are paying for a conversion action: a purchase, a sign-up, a phone call, a booking, whichever business event you decide has value. Naming that action is the first task, and it belongs to you rather than the provider, because only you know whether a form submission from forty miles away is a lead or a nuisance. Write down the one or two actions that count and what makes one qualified.

Then take the baseline. For the three months before anything starts, record revenue or gross margin if you have it, every marketing cost by the five lines above, the number of enquiries, how many were qualified, how many booked, and where each one said they came from. A platform report is not that record. Your inbox, your calendar and your invoices are, and the person best placed to keep it is the one who answers the phone.

With the action defined and the baseline kept, the six months have a shape. Month one is for verifying that the tracking actually captures the action you named, and that you have access to the accounts where it is reported. Months one to six are for the ledger: each month, the actual spend by line, the conversion count from the platform, the qualified-lead count from your own records, booked or closed value where you have it, and a note of anything that changed. At six months, you compare the same definitions over the same length of period against the baseline.

The measures themselves are worth knowing by name so a report cannot blur them. Google Ads defines cost per conversion as total cost divided by conversions, and conversion value per cost as total conversion value divided by the cost of ad interactions. Search Console's performance report gives clicks, impressions, click-through rate and average position for organic search. A verified Google Business Profile reports calls, website clicks, direction requests, messages and bookings. Every one of those is a platform measure of attention or action. None is a sale until your own record connects it to one.

The change log is the part most often skipped and most often missed later. Without it, a rise or fall in the sixth month cannot be separated from the price change in March, the tracking fix in April, the week you were closed, or the season. It is one dated line per change, and it takes a minute.

Six months is long enough to see whether enquiries moved, whether the ones that came were the kind you wanted, and what each cost. It is long enough to see whether the provider did what the scope said, on the cadence agreed, with the changes logged. Nothing about six months makes it a guaranteed proof period, and a provider who calls it one is offering a comfort rather than a fact.

The 7 to 14 day figure that appears in Google's campaign guidance is a waiting period after launch or after major changes to bids, creative or targeting, given in the context of a particular campaign type with limited conversion history. Its use to you is as a reason not to change a campaign weekly, since Google warns that repeated changes force the bidding to relearn. Google also lets a conversion window run from 1 to 90 days depending on the source, and recommends at least 7, so a business with a long buying cycle should pick a window that matches it or the later conversions vanish from the count.

What six months of clicks, impressions and profile views cannot tell you is whether the business made money from them. That answer needs the qualified-lead count and the booked value from your own records, and if those two columns are empty at month six, the review can establish attention and not much else. Keep the conclusion at the level the records support: continue, revise one defined element, pause, or gather better data.

The scope, the fees and the deliverables are the terms everybody reads. The four that decide what the six-month review is worth are read less often.

Who funds the advertising. If media spend runs through the provider's account on the provider's card, the platform history, the conversion definitions and the learned bidding data are in their account. Google's own documentation makes conversion configuration material to what the reports show and how bidding behaves, which means an account you cannot take with you is a six-month record you cannot take with you either. Have the ad account, the analytics, the Business Profile and the website in your name, with the provider as a user you can remove.

What survives termination. Access, exports, the creative, and the historical reports. Ask what happens on the day after, in writing.

Renewal and cancellation. The Texas State Law Library's consumer FAQ says Texas has no statute regulating auto-renewing contracts and that contract law is complex, which for you means the renewal clause says whatever it says and there is no statute behind you if you missed it. Read the notice period, the renewal term and the exit mechanism before the term begins.

Who approves the claims. The FTC's advertising guidance says an advertiser needs a reasonable basis, meaning objective evidence, for claims, and that endorsements must reflect honest experience. Texas Business and Commerce Code Section 17.46 makes false, misleading or deceptive practices in trade unlawful. The advertiser in both is you, not the person who drafted the ad. A contract that lets copy go live without your sign-off is a contract that lets somebody else's sentence become your liability. This is not legal advice, and an agreement with meaningful liability, ownership or dispute terms is worth a Texas lawyer's hour.

An owner can do the foundational work without help and should, because it is the part that makes any later spend judgeable. Deciding the objective and the valuable action. Keeping the spend and lead ledger. Marking which enquiries were qualified and which booked. Preserving invoices. Keeping the hours, services and area facts current everywhere they appear. Reading your own Search Console and Business Profile reports, which are built to be read by owners.

The line to stop at is the one where a mistake is expensive or invisible. Changing site code without a backup. Handing account credentials to anyone whose identity and role you have not confirmed. Publishing a performance claim or a comparison without evidence behind it. Running a testimonial that does not reflect a real customer's honest experience. Setting up tracking that collects visitor data without understanding the consent it needs, which Google's own policies require where the law does. Each of those is a place to get the right help rather than push on.

If you do the work yourself, record your hours in the ledger the same way you would record a fee, so the comparison with a paid option stays honest. Owner time is not free. It is just unpriced until you price it.

None of them sets a marketing budget, but several change what an Arlington business has available before it spends. The city's small-business resources page describes the Arlington Economic Development Corporation's industry-specific small-business cohorts, 0% interest loans through LiftFund, and a Strategic Solutions Fund for Arlington-based manufacturers. It also points to the Greater Arlington Chamber's Keep it Local directory, a listing that costs nothing to be accurate in.

Businesses inside the Downtown Arlington Business Improvement District boundary sit in a self-assessed programme that Downtown Arlington Management Corporation says covers marketing, economic development, security and beautification under a city contract, with a storefront-improvement grant for eligible physical work such as signage, awnings, paint and lighting. It is a place programme rather than a marketing subsidy, and it applies only inside the boundary, so check the map before counting on it.

What the local picture does not contain is a published benchmark for what marketing help costs in Arlington. The number is whatever the proposals say once they are split into their five parts.

The whole framework fits on two pages, and it outlives the provider, the platform and the person who kept it. Page one, written at the start: the budget cap and its five components, the valuable action and what makes a lead qualified, the baseline period and its figures, where the real lead and sales figures live, who owns each account, the review dates, and the rule for approving a change. Page two, one line a month: spend by line, conversions, qualified leads, booked value, and what changed.

At month six the review is arithmetic on those two pages, done with the same definitions over the same length of time. Cost per qualified lead from your own ledger. Cost per conversion from the platform where you ran ads. Clicks and positions from Search Console as context for search visibility. Profile actions as context for local interest. And a conclusion no bigger than the evidence: keep going, change one thing, stop, or measure better.

A better business result is not guaranteed by keeping the record. A defensible decision is, and that is the thing you did not have when the three proposals landed.

What percentage of revenue should a small business spend on marketing?

The published percentages come from surveys of other populations: BDC's 2% to 10% is from 2012, Canadian and combines sales with marketing; Gartner's 7.7% is from marketing leaders at mostly billion-dollar companies. They give a sense of scale. The right number for your business comes from your revenue, your margin per job and what a customer is worth over time, tested against your own records.

How much do small businesses spend on marketing?

Intuit's 2025 report estimates an average small-business advertising budget of $78,000, with advertising averaging 37% of the total marketing budget and labour excluded. That is a national estimate about advertising, not a figure about what a retainer costs or what your business should spend.

How much should I spend on digital marketing versus everything else?

Decide the valuable action first, then fund the channels most likely to produce it, and let the six-month ledger tell you where the qualified leads actually came from. Any fixed split chosen before that evidence exists is a guess, however tidy.

Is six months long enough to know if it worked?

Long enough to see whether enquiries moved and what they cost, if the baseline and the ledger were kept. Six months is a review point rather than a proof period, and Google's only stated waiting figure is 7 to 14 days after major paid-campaign changes in a specific context.

Should the agency own the ad account?

It is operationally convenient for them and costly for you if you part ways, because the history, conversion setup and learned bidding stay in their account. Hold the account in your name and give them access you can remove.

What if the quotes include ad spend and I cannot tell how much?

Ask for the five-line breakdown: fee, media, software, one-time work, and what is expected of you. A proposal that cannot be broken out that way cannot be compared with one that can.

The percentage you were looking for was never the answer, because every published one describes a population you are not in. The answer is a number you can take apart into its five pieces, compare across proposals, and check against your own ledger at month six.

That check is the part worth protecting. A baseline kept before the spend starts, an action defined before the first report arrives, and a change log kept through the middle turn a renewal decision from a feeling into arithmetic.

If your proposals include a website or search work and you want a second pair of eyes on what is actually in them, Arlington Website Designer reads proposals for owners in Arlington, TX and will tell you plainly which of the five parts each one covers. Send the quotes, with names removed if you prefer, through the contact page, and you will get back a like-for-like breakdown and the questions still worth asking each provider.