An owner once asked us, in the middle of a proposal meeting, to walk them through a time our process failed a client. Then they asked how we’d failed our customers, because that would tell them a lot about who we were. It’s still the best question I’ve heard from a buyer.
I’ve run a marketing agency, Connective Web Design, since 2010, so read this knowing I have a financial interest whenever a company like yours hires an agency like mine. I’ve also sat on the other side of the table, leading marketing for companies that hired agencies and needed someone to manage them, and I’ve watched what happens when the wrong one gets hired. And I’ve been on the agency side when what sales promised got ahead of what the delivery team could build, and my company paid for it. Further down is the question that would have caught it.
The short answer to how you choose a marketing agency, digital or otherwise: work out your own numbers first, decide what kind of help you actually need, talk to about three agencies whose relevant work you can see, ask all of them the same questions, check their answers against evidence, compare the proposals line by line, and make sure you own everything before you sign. Then agree on a date to judge the results, and give the plan a real chance to work before you do.
Most of choosing well comes down to two things: knowing what an outcome is worth to you, and getting evidence that this particular team can produce it, on terms that let you recover if they can’t.
First, decide whether you need an agency at all
An agency is the right answer when you need several skills at once, at a volume you can’t justify hiring for, and you want someone else to manage the people doing the work. It’s often the wrong answer in a few situations:
- The work is one narrow specialty. A good freelance specialist or a smaller boutique may do it better and cheaper.
- Marketing is central to how you grow and the volume is steady. An in-house hire, with outside help for the specialist pieces, may serve you better over time.
- Nobody has decided what marketing should achieve. Then you need the thinking before you need the hands. That’s a strategist, a fractional CMO or an advisor, not a production team.
- The budget can’t support the goal. If the campaign needs more money than the business can carry, not doing it is a legitimate choice.
Over the years my own agency has talked companies out of working with us, and helped clients build their own in-house teams knowing it meant less work for us. It’s worth asking any agency you talk to whether they’ve ever done the same, and what happened.
Whoever you hire, someone on your side still has to steer. A few of the companies I’ve led marketing for hired me partly to be their voice with their own vendors. One of those vendors was a paid-search specialist running very large ad budgets. They knew the tools cold, and they were optimizing the account toward numbers that didn’t connect closely enough to what the business needed. Nobody on the client side had the background to challenge them, so nobody did. Specialists are worth hiring. Somebody owner-minded has to point them at the business.
Work out your own numbers before the first call
I run a marketing agency, and I don’t think you should hand your marketing goal-setting entirely to your agency. You know what a customer is worth to you better than they do.
The most useful number to bring is how much you could spend to win one new customer. A simple way to estimate it, with made-up numbers:
- Say you sell a service for $100,000.
- Delivering it costs $60,000, which leaves $40,000.
- Every project carries its share of overhead. Say that’s $15,000, which leaves $25,000.
- A business isn’t really a business unless it’s profitable, so back out a minimum profit. Say $10,000.
- That leaves $15,000: the most you’d want to spend on sales and marketing combined to win that customer.
- Part of that goes to selling: your salespeople’s time, commissions, travel, the proposal itself. Say that’s $5,000 per customer you win, which leaves about $10,000 for marketing.
If your customers buy again, use the profit over the whole relationship, not just the first sale. Then turn the marketing share into a lead number: if you close one in four qualified leads, $10,000 means you can afford about $2,500 to produce each qualified lead. That’s a number an agency can actually work against.
It’s a directional ceiling under your own assumptions, not a law, and attribution gets messy in the real world. But walk in with it and you have something you and the agency can work toward beating, month after month. Walk in without it and the agency ends up setting the target for you.
Bring a few other things too:
- A budget range you can commit to for at least a few months.
- What you’ve already tried, and what happened.
- Who on your side owns the relationship and can make decisions quickly.
- What you can measure today. If you can’t tell where your good customers come from now, fixing that may be the first job.
One question worth asking yourself before you ask any agency: are you expecting $75,000 results on a $2,000 budget? If the answer is yes, no agency will fix that, and the good ones will tell you so.
Know what kind of agency you’re looking for
Whether you’re choosing a digital marketing agency or a broader one, most agencies fall into a few shapes, and the right one depends on the problem.
| Type | Good fit when | Watch for |
|---|---|---|
| Specialist (SEO, paid search, social, PR) | You know which channel matters and need depth in it | Optimizing their channel without looking at the rest of the business |
| Creative or brand agency | Your positioning, identity or messaging needs work | Beautiful work that nobody connected to sales |
| Web design and development | You need a new site or a serious rebuild | Who owns the site afterward, and who can change it |
| Integrated or “growth” agency | You need several channels working from one plan | Being average at everything, with your account split across junior staff |
| Freelancer or small boutique | The scope is narrow, or budget is tight | Capacity, and what happens when that one person is unavailable |
Size matters less than fit. The question is whether your account will matter to them. At a large agency, a small account can end up with the most junior people. At a very small one, the risk is capacity.
How marketing agencies charge
Pricing varies too much by scope to give you a single number, and I’d be careful with anyone who quotes one before they understand your business. What helps more is knowing the models and what each one rewards.
| Model | How it works | What it rewards | Ask |
|---|---|---|---|
| Monthly retainer | A fixed fee for a defined level of ongoing work | Steady attention | What the fee buys each month, and what happens to unused time |
| Project fee | A fixed price for a defined deliverable | Finishing the defined scope | What’s excluded, and how changes get priced |
| Hourly | You pay for time used | Accuracy on unpredictable work | How hours are approved and reported |
| Percentage of ad spend | The fee rises and falls with your media budget | Spending more | Whether the fee tracks the work or just the budget |
| Performance-based | The fee depends on results | Results they can claim credit for | How results are measured, and who decides |
When two quotes are far apart, they’re usually describing different amounts of work. In a website cost breakdown I put together, two fair proposals for what sounded like the same site came in at very different prices, and each was a reasonable price for what it included. Ask every agency what levers drive their price, and what the extra money buys. A good one can explain the trade-offs. Then you can decide whether you need them. Sometimes the more affordable option is the right call.
Build a short list of about three
For most privately held companies, three serious candidates is enough to show you real differences without turning the search into a procurement exercise. Four can make sense if their approaches are very different. I wouldn’t add agencies just to reach a number.
Start with referrals from people whose results you trust, then look at agencies whose work you’ve actually seen in your industry or at your scale. Directories and award lists are fine for finding names, but they don’t tell you much about fit. Familiarity alone isn’t a qualification, and that includes mine.
Before you invest time, ask each one whether they have the capacity to start when you need them, and who would actually be on your account.
Give the whole search a few weeks, not a few days: one week to prepare, a week or two of first conversations, then proposals, reference calls and a decision.
Write a one-page brief instead of a long RFP
A short, specific brief gets you better proposals than a forty-page marketing agency RFP. Plenty of strong agencies pass on RFPs that ask for free strategy and hide the budget, and a long RFP tends to favor the agencies with the most sales capacity, not the ones best suited to the work.
A useful brief fits on one page:
- The business: what you sell, to whom, and roughly how big you are
- The problem: what isn’t working, in your own words
- The goal: what result would make this worth it, and by when
- Your numbers: what a customer is worth, and your acquisition-cost ceiling
- Budget range: what you can commit to, and for how long
- What you’ve tried: and what happened
- Who decides: names, and how quickly decisions get made
- What’s in place: your website platform, CRM, analytics and any current vendors
- Timeline: when you need to start, and when you’ll decide
Share the budget range. Without it, you’ll get proposals sized to guesses, and you’ll waste everyone’s time comparing them.
A formal RFP does make sense in some cases: when procurement requires one, when the spend is large, or when you’re coordinating several vendors at once. Even then, keep it short, pay for any real strategy work you ask for, and tell the agencies how you’ll score them.
The questions to ask a marketing agency, and what good answers sound like
Ask every agency the same questions so you can compare the answers. The answers matter less than how specific they are and whether you can check them.
| Question | A good answer | A warning sign |
|---|---|---|
| At my budget, what can I realistically expect, and by when? | Specific, a little cautious, and names its assumptions | A confident promise of rankings, leads or revenue by a date |
| What would make us change course? | Names the data that would tell you it isn’t working | “We’ll keep optimizing” |
| Who will actually do the work, and how much of their time is mine? | Names and roles you can meet | “Our team” with no names |
| Which person here is the reason we’d succeed with you, and how involved will they be six months from now? | A name, a role and a realistic answer about their time | The founder in the pitch, then never again |
| Has the person who’ll deliver this seen what’s being promised? | Yes, and they can show you how sales and delivery agree on scope | Surprise at the question |
| What drives your price? | Explains the levers and trade-offs in plain language | “That’s our standard package” |
| What’s not included, and how are changes priced? | A clear list, and a clear change process | “We’re flexible” |
| Can you walk me through a time a client wasn’t happy? | A real story: what went wrong, what they did, what changed | “That doesn’t really happen to us” |
| Have you ever told a company not to hire you? | Yes, with an example and the reason | A pause, then a pitch |
| What will you need from us, and how fast? | Specific asks, approvals and turnaround times | “Not much, we handle everything” |
| Who owns the accounts, the site and the data? | You do, in your name, from day one | Accounts set up under the agency |
| What does month one look like? | A concrete plan: access, baseline, first priorities | A long list of tactics |
| How do you report, and what do you measure? | Business outcomes, with a sample report you can see | Impressions and traffic, with no line to sales |
| How does this end, if it ends? | Notice period, handover and what you keep | Silence, or a long lock-in |
The “client wasn’t happy” question is the one that owner asked us, and it’s the best one on this list. Every agency has a story like that. The ones worth hiring will tell it.
Check the claims before you believe them
A capable salesperson can pass every question in that table. The step most people skip is checking the answers against evidence.
- Relevant work. Ask for two or three examples close to your situation, and ask what the agency did versus what the client’s team or other vendors did.
- References. Every reference an agency hands you will like them, so “are they good?” won’t tell you much. Ask what they’re worse at than you expected, when you have to manage them, who on their team matters most, and what kind of company shouldn’t hire them. If they’ll give you a former client, take it.
- A sample report, with the client’s details removed. It shows you what they think matters.
- Their own marketing. If they sell SEO, look at how they show up in search. It isn’t everything, but it’s evidence.
- Your sales lead, in the room. Bring whoever runs sales to at least one meeting. How an agency answers their questions about lead quality tells you whether it thinks past the click.
That last one comes from a mistake I’ve made. On a sales call, a salesperson once told a client we’d build “the best thing on the internet.” What they meant was that the presentation of a tool would be better. The plan was to use a third-party tool behind it. The client heard a custom-built program unlike anything else online. By the time the gap surfaced, the project was well along, and we ended up paying to do the work the client thought they’d bought. Now our delivery team signs off on what sales promises before anything is signed: that we can produce it on time, on budget and at the expected quality, at the price stated. Ask any agency how they handle that gap. Then ask to meet the people who’ll be doing the work, and ask whether they’ve seen the proposal.
Look for evidence that matches your problem
Competence is the easier thing to check. The harder question is whether this agency can solve your particular problem. Weigh the evidence roughly in this order:
- The same business problem matters more than the same industry.
- The same economics and sales motion. A six-month B2B sale and a same-day online purchase are different jobs.
- The same channel, when one channel carries most of the weight.
- A similar company size, because budgets, approvals and internal teams change what’s possible.
A logo in a portfolio tells you very little until you know what the agency actually did for it. If you’re a $20 million manufacturer trying to win qualified opportunities from a six-month sales cycle, an ecommerce traffic win doesn’t tell you much, even if that company makes things too.
How to compare proposals that look nothing alike
Every agency packages its work differently, which is why proposals are hard to compare. Put them side by side in a spreadsheet: one column per agency, one row per thing that matters.
| Compare | What to look for |
|---|---|
| Diagnosis | Did they understand your business, or could this proposal go to anyone? |
| Deliverables | What you get, how many, how often |
| Exclusions | What’s out of scope, and what it costs to add |
| Staffing | Named people, their roles, and how much of their time you get |
| Your responsibilities | What you have to supply, approve or attend, and how quickly |
| Measures and review dates | What success looks like, and when you’ll check |
| Total cost | Fees, plus ad spend, software, licenses and other outside costs |
| Exit | Notice period, handover, and what you keep |
If I could only read one row, it would be the diagnosis. A proposal that could have gone to any company usually means the thinking hasn’t happened yet. If it helps, score each row from 1 to 5 and weight the rows that matter most to you. The total is less useful than the conversation it forces. When one proposal costs far more, the grid shows you exactly where the money goes, and you can decide whether you need that part.
Two details catch people out:
- Revisions. An agency that includes two rounds of revisions on everything is probably pricing in that you’ll use them. Ask how they treat corrections versus changes. A correction is when something doesn’t match what you already approved, and that’s on the agency. A change is when you ask for something different, and that usually means more cost.
- Contingency. Whichever way you go, set aside 5 to 10% of the project price for the things that come up along the way.
What to settle before you sign
A lot of the agency disputes I’ve seen started with assumptions nobody wrote down. Confirm these before you sign, while everyone is friendly:
- You own your accounts. Ad accounts, analytics, your domain, your hosting and your social profiles should be in your company’s name, with the agency given access.
- You own the work. Which files, designs, code and content you receive, and when.
- What doesn’t transfer. Licensed images, fonts, plugins and third-party tools may not come with you, and some keep costing money after the relationship ends.
- The term and the way out. How long you’re committed, how much notice ends it, and what handover looks like.
- How changes are priced. The process for anything outside the agreed scope.
- Who approves what. On your side and theirs, and how quickly. An agency can only move as fast as your approvals.
- What happens if a key person leaves. For a large account or critical systems, ask how they cover turnover.
- How sales feedback reaches them. If your sales team knows which leads turned into customers, make sure that gets back to the agency. It’s one of the most useful things you can give them.
If you take one thing from this section: never let the agency hold the keys. I’ve seen companies held hostage by a vendor who owned their hosting or their ad account. Own your own stuff, and give the agency access to it.
For anything beyond these practical checks, have someone qualified review the contract itself.
Red flags
Most of the warning signs before you sign are in the questions table above. Two more are worth naming on their own: guaranteed rankings, leads or revenue by a specific date, and tactics recommended before anyone has asked about your business, your margins or your best customers.
After you sign, watch for:
- Communication fades once the contract is in place
- Reports keep getting longer while results stay the same
- You can’t get a clear answer about what’s working and what isn’t
- Commitments keep slipping without an explanation
Limit what you can’t undo
If you’re unsure, reduce how much you commit before you have evidence. Sometimes that’s a paid discovery phase. Sometimes it’s a shorter first term, a defined first milestone, or a 90-day review with a way out. Making the first commitment reversible matters more than making it small. Some work, like SEO or a full website, doesn’t split neatly into a three-month trial, and a polished audit doesn’t prove an agency can execute.
If an agency requires a long first commitment, ask why. There can be a good reason. What matters is what they have to invest before results can fairly be judged, and what protects you if delivery falls short.
The first 90 days, and how to judge them
Agree on a review date before the work starts, and agree on what you’ll look at when it comes. Then let the plan run.
Be clear about which numbers belong to whom. A paid-search agency controls targeting, bids, ads and tracking. It influences how many qualified leads you get. It usually doesn’t control whether your sales team calls those leads back, or closes them. I want an agency that understands revenue, but holding it to a number it can’t move mostly teaches everyone to argue about attribution. Agree up front on what they own, what you own, and what you’re moving together.
When the date comes, there are three real decisions: keep going, change course, or stop. Stopping is a legitimate decision, especially if the budget can’t support the goal.
Some things shouldn’t wait for the review date: work that’s plainly wrong, tracking that’s broken, commitments that keep getting missed, or a business that’s changed underneath the plan.
And before you replace an agency, ask which of these it was: the idea was weak, the execution was poor, your side held things up, or conditions changed. Only some of those are fixed by switching. If nobody set expectations and review dates, look at the agency first. If they did, and the plan keeps changing, look at your side.
If you’re replacing an agency you already have
Most owners reading this already have an agency, so a few things to do before you give notice:
- Get your access back first. Make sure you hold administrator access to your ad accounts, analytics, website, domain and hosting before the conversation, not after.
- Ask for the history. Campaign settings, past reports, keyword and audience lists, creative files and any tracking setup. It’s much easier to ask while the relationship is still working.
- Overlap if you can. A few weeks where the new agency learns the accounts while the old one is still in place beats a hard cutover.
- Keep it civil. You may need their help during the handover, and how they handle the ending tells you a lot about how they ran the work.
If you’re choosing a web design agency
Everything above applies. The biggest difference between website quotes is usually who does the production work. I’ve seen two quotes for what sounded like the same site where one agency designed a few templates and left the client to build out the pages, and the other did all of it. Neither was wrong. They were different projects, and only one of them matched what the client had in mind.
So ask each one how much of the finished site they’ll build and how much they expect from you, and three more things:
- Who owns the site when it’s done, who holds every login, and what it costs to keep running.
- Whether your team can make everyday changes without them, or whether every small edit goes back through the agency.
- How they’ll protect your search rankings if you’re moving from an existing site.
I wrote a fuller breakdown of how to compare website quotes in the post on how much a website should cost.
Common questions
How long before an agency shows results?
It depends on the channel. Paid search can show early signals within weeks. SEO, content and brand work usually take months. Anyone promising specific results by a specific date is probably overpromising.
Should I hire a marketing agency or someone in-house?
An agency gives you several skills at once without hiring for each. An in-house hire makes more sense when marketing is central to how you grow and the work is steady. Many companies end up with both: a strong internal lead and outside help for the specialist work.
A last word on choosing
If I were spending my own money, I’d rank the decision roughly this way. First, do I believe their diagnosis? Second, have they shown evidence they can solve a problem like mine? Third, do I trust the people who’ll actually do the work? Fourth, can we agree on how success gets measured? Only then scope and price. I wouldn’t pick the cheapest agency, and I wouldn’t assume the most expensive one is better. I’d pick the least expensive option I genuinely believe can solve the problem.
That takes a few weeks of work on your side. It’s a lot cheaper than finding out a year in.
If you’d like someone on your side of the table while you do it, I help owners run agency searches: the brief, the questions, the reference checks and a side-by-side read of the proposals. Connective is on the list only if you want it there. If it is, I have an obvious conflict, so weigh my opinion of it accordingly: it goes through the same questions as everyone else, and you make the decision. If you’d rather run a search with no conflict at all, say so, and it won’t be considered. You can see how that works on the agency hiring advisor page, or start a conversation on Work With Me.
