Paid Media
We've had machine learning making bid decisions since 2013.
Our media team plans and runs paid search, paid social, programmatic display, connected TV, streaming audio, and digital out-of-home, and we hold direct seats on the demand-side platforms, so the channel recommendation isn't shaped by inventory we need to move. Some clients hire us to run the whole program. Some hire us for a single audit of an account they inherited and can't make sense of. Both are a good way to work with us.
One program, six channels
Thirteen years of running this in production, and eighty-seven people deciding what it's allowed to do.
We started putting machine learning into bid and budget management in 2013, which is long enough ago that we can be specific about where it helps and where it creates risk you don't want. The system works on top of the platforms' own automation, and a person still decides what it's allowed to touch.
2013
AI in production on bid and budget
Live bid and budget decisions across every major platform, running since then.
30 min
Optimization cadence
How often the system reacts to live auction behavior.
11
Platforms we buy on
Google, Microsoft, Meta, LinkedIn, TikTok, Pinterest, Reddit, Yahoo DSP, Simpli.fi, Illumin, and OpenAI Ads.
Selected clients
Four situations bring most people to this page.
01
Every channel is winning and the business isn't.
The search report is up, the social dashboard is green, the programmatic vendor sent a deck full of impressions, and revenue is flat. This is what happens when each channel optimizes the number it owns and nobody answers for the number that matters. It's the most common thing we get hired to untangle, and it's usually a measurement problem wearing a media costume.
02
The whole program is search, and search has run out of room.
Non-brand search captures demand that already exists, so once you own the terms worth owning, more budget buys progressively worse clicks. We see this most in categories where the paid search account has been run well for years and the growth target went up anyway. The next increment has to come from channels that create demand, and those channels need a different plan and a different way of reading the results.
03
You inherited an account you can't audit.
The agency owns the ad accounts, the reporting arrives as a monthly PDF, and nobody can tell you what changed last Tuesday or why. We take over mature programs regularly, and the first thing we deliver is an honest read on what the previous team was actually doing, including the parts that were working.
04
The conversions stopped being countable.
Cookie loss, in-app browsers, form fills that never become revenue, and platform-reported conversions that add up to more than the sales you actually made. We'll say the uncomfortable version: this happened to our own program. A landing platform returned its success message inline on the same URL, and a conversion action that required a page load recorded nothing for ninety days while leads were arriving the whole time. Every marketer reading this has lived some version of it, and it's fixable in about a week once somebody looks.
Our paid media approach is art and science, in equal measure.
Growth doesn't come from running ads. It comes from a program where every dollar traces back to a business goal, and where the strategy, the creative, the audience, and the measurement each do their job. We've built ours around two halves that have to work together.
The art
Strategy and creative judgment that machines can't replace.
Audience and persona work. Who actually buys, and why.
Channel and platform strategy. Where attention is, and not simply where the impressions are cheap.
Creative direction. Message-market fit across formats, tested against each other and not against an internal opinion.
Funnel architecture. Sequencing prospecting, nurture, and conversion so the channels stop competing for the same click.
The science
Our own technology and measurement, deciding where each dollar goes.
AI on bid and budget. In production since 2013, across every major platform.
A thirty-minute cadence. The system reads live auction behavior and adjusts, without waiting for a weekly optimization pass.
Cross-channel attribution and incrementality testing. Covered in detail further down this page.
Consolidated reporting. One view that ties spend to revenue, updating through the day.
The two halves are the whole argument. Automation without judgment optimizes toward whatever is easiest to measure, and judgment without automation can't react fast enough to matter.
We'll tell you which decisions our system makes on its own and which ones a person still signs off on, because in a regulated category that distinction is the entire conversation.
Let's have a chat →Media is a lot more than search now, and we run all of it end to end.
We don't subcontract media. Every channel below is staffed, certified, and active in the agency today, which means strategy, build, optimization, and reporting all happen in the same room.
01
Paid search
Branded, non-brand, Shopping, and Performance Max.
02
Paid social
03
Demand generation
Upper funnel at scale.
04
Programmatic display
Audience, contextual, and addressable targeting, bought on our own seats.
05
CTV and streaming
Premium inventory, bought against the streaming apps people actually watch.
06
Streaming audio
Ad-supported music and podcasts.
07
Retargeting and retention
Cross-channel sequencing.
08
Emerging surfaces
Pixels inside AI answer surfaces.
05a · The seats, and what they get you
We hold direct seats, which is why the recommendation can be honest.
We have no inventory to move and no volume commitment to hit, so when we tell you programmatic is wrong for what you're trying to do this quarter, that recommendation costs us money. We map each platform to the job it's good at.
Your budget
media budget
Platform minimum
Bypassed
Our seats
Scale · premium CTV
Addressable geo
Journey canvas
Premium inventory
The seats are also how you get access at all.
The demand-side platforms price access for buyers at national scale, with platform minimums and annual commitments that put premium programmatic out of reach for most of the companies that would benefit from it. Because your campaigns run on our seats, you get the same premium CTV inventory, the same identity graph, and the same addressable targeting a national advertiser gets, without signing a platform commitment to reach it. This is the one place where our infrastructure genuinely is the product.
What you give up by not holding the seat is nothing you'd notice, because the visibility comes with it: line-level delivery detail, the inventory reports, the site and app lists, the seller IDs, and the exclusions we've applied on your behalf. The full ownership model is spelled out further down this page.
We're already running pixels inside AI answer surfaces.
Our media team has built OpenAI Ads pixels on two client accounts, one for a global hospitality brand's franchise development program and one for a direct-to-consumer eyewear retailer. That puts us among the first agencies instrumenting measurement inside an AI assistant instead of writing about the possibility of it.
Here's the honest read. The inventory is new, the volume is small, the reporting is thinner than any established channel, and we're treating the whole thing as a measured test with a defined budget, and not as a line item we're selling into plans. What it buys you today is a baseline and a working pixel, so that when the volume arrives you have a year of your own data instead of a vendor's case study. If your buyers are already asking assistants about your category, this is worth a small, instrumented test now, and it isn't worth a large one.
The work covers six areas, and you can engage us on one of them or all of them.
Most programs pull from several. Each phase is scoped and priced on its own, and a one-time audit with a prioritized roadmap is a perfectly good way to work with us.
01Program strategy and media planningAudience, offer, channel mix, budget allocation, and the forecast you'll be held to.+
Our media strategist starts with what you're actually trying to buy, whether that's qualified pipeline, applications, first orders, or repeat purchase rate, and works backward into a channel mix and a budget allocation that can hit it. That includes a forecast built bottom-up from your own historical performance and current auction conditions, with the assumptions written down, and it includes the funnel architecture that says what each channel is responsible for, so prospecting and retargeting stop taking credit for each other's work. Where the honest answer is that your target isn't reachable on the budget available, we'll show you the math instead of taking the assignment.
02Build and activationCampaign architecture, audience build, feed and catalog work, creative trafficking, and the tagging that comes first.+
Our specialists build the account structure, the keyword and audience sets, the geographic and dayparting logic, and the creative trafficking across every platform in the program. On commerce accounts, feed and catalog quality gets treated as part of the build, because a product feed with missing attributes caps Shopping and Performance Max performance in ways no amount of bidding fixes. Nothing spends until conversion tracking is verified end to end, which is a rule and not a preference, and the reason for it is in the next band.
03AI-led bid and budget managementOur own system, in production since 2013, running on top of platform automation.+
Our technology makes live bid and budget decisions across Google, Microsoft, Meta, LinkedIn, TikTok, Yahoo, Reddit, and Pinterest, reacting to auction behavior as often as every thirty minutes. It layers on top of each platform's native automation, which matters because Smart Bidding and Advantage+ are genuinely good at the things they're good at, and the value we add sits in budget allocation across campaigns and channels, target setting, and catching the drift that platform automation introduces when the conversion signal it's optimizing toward is the wrong one.
04Creative production and testingConcepting, direction, production across formats, and fatigue monitoring.+
Our creative group concepts and produces across every format the program needs, from responsive search assets through display, video, CTV spots, and audio. Testing runs as a structured program with a defined hypothesis behind each variant, and we monitor creative fatigue as its own signal, because a set of assets usually stops persuading people several weeks before the platform's efficiency metrics register the decline. You get told when a creative is done, along with what we'd replace it with.
05Measurement, attribution, and conversion trackingGA4 and GTM implementation, server-side, offline conversion import, and incrementality testing.+
Our analytics team owns this work, and it's substantial enough that it has its own band further down the page.
06Audits and account takeoversA fixed-fee audit with a prioritized roadmap, on an account you already own.+
We audit accounts we didn't build, which covers the campaign structure, the bidding configuration, the audience and exclusion logic, the creative, the landing pages, and, most often the actual problem, the conversion tracking. You get a written findings document, a prioritized roadmap, and a working call to go through it. Plenty of clients stop there, and the roadmap is written to be handed to whoever ends up doing the work, including your in-house team or another agency.
Runs through all six
AI
AI isn't a seventh service here. It's a decision inside each of the six, and the useful question is always which part of a workflow it should touch. We've been running it in production since 2013, which is why we're specific about the answer.
Connects all six
Measurement
Measurement is what lets the other five be argued about honestly. If the numbers can't be trusted, no recommendation can be checked and every channel gets to claim the win.
Every channel will claim the same conversion, so somebody has to arbitrate.
A form fill is not a sale, and optimizing toward one gets you more of the wrong thing.
In a business with a six-month cycle and a buying committee, the leads that are easiest to collect and the leads that become customers are two different populations, and the bidding algorithms can't tell them apart unless you tell them. Our analytics team wires closed-won outcomes from your CRM back into the ad platforms, so the optimization runs against revenue and not against inquiry volume. The practical effect is that the campaign generating a pile of cheap, unqualified form fills gets defunded by the algorithm itself, without anyone having to argue about lead quality on a monthly call. For B2B programs this is the requirement and not the upgrade, and it's the first thing we build.
The same logic applies wherever the outcome that matters happens somewhere other than your website. In higher education we've been joining Slate inquiry records to behavioral data since 2013, so admissions teams can see which campaigns and keywords produced applicants and deposits, not sessions.
The rest of the measurement stack.
Our analytics team handles GA4 and GTM implementation, server-side migrations, and consent-mode configuration, which is increasingly where the accuracy of everything downstream is decided. On top of that we run multi-touch attribution where the data supports it and media mix modeling where the spend justifies the exercise, and we'll tell you when it doesn't, because an MMM built on eight months of thin data produces confident numbers that aren't true. Incrementality testing and geo holdouts get used where a channel's contribution is genuinely in question, since withholding a channel in a matched market and watching what happens is the only method that answers causation.
What we won't do is hand you a number and ask you to trust it.
Attribution is a set of trade-offs, not a truth machine. Every model makes assumptions about credit, windows, and what counts as a touch, and those assumptions change the answer. We show you the methodology and the assumptions alongside the number, so you can decide whether you believe it. No black boxes, and no convenient math.
You see what we see, in near real time.
Most agency reporting is a monthly artifact assembled by the people whose work it evaluates. Ours is a live view of the program that you have access to on the same terms we do.
Drawn rather than captured, since the live view carries client spend.
One view across every platform. Search, social, programmatic, CTV, audio, and DOOH in a single place, with conversions defined the same way across all of them. No separate login and no separate definition per channel.
Spend and pacing that update through the day. A budget problem surfaces the day it happens instead of in a month-end recap, which is the difference between a conversation about a fix and a conversation about a refund.
Creative-level detail. Which asset is carrying the account, and which one has quietly stopped working.
Filtering that follows your business. Campaign, campaign group, geography, and audience, cut the way you actually think about the program, and not the way a reporting tool prefers to organize it.
Access as standard. Not an upgrade, not a quarterly business review deliverable, and not a screenshot pasted into a deck.
Reporting that only flatters the agency isn't reporting.
The reason we can offer this is that the underlying numbers hold up, and the reason they hold up is the measurement work in the band above.
Let's have a chat →Here's how a media engagement actually runs.
We won't launch a campaign until conversion tracking is verified end to end. That sounds obvious, and it's the single most common reason a program looks like it's failing when it isn't.
01 · WEEKS 1 TO 3
Audit and baseline
An account audit, a tracking audit, a landing page review, and a measurement baseline, followed by a channel and budget recommendation and a prioritized roadmap. You get a written findings document and a working call to go through it. Plenty of clients stop here.
02 · WEEKS 2 TO 6
Instrument and build
Our analytics team wires the tags, conversion actions, and offline import, and QAs them against real submissions before anything spends. In parallel, our specialists build the campaign architecture, the audiences, and the creative against the plan.
03 · WEEKS 5 TO 10
Launch and learn
Campaigns go live with the test structure defined up front, so the learning period produces answers. Budgets ramp deliberately, because a platform's learning phase resets every time you make a large change and the first two weeks of data on a new channel are the least reliable data you'll ever get from it.
04 · MONTHLY, ONGOING
Optimize and report
Bid and budget management runs continuously in the background. On top of that, our team refreshes creative on a defined cadence, runs incrementality tests where they'll settle something, and reports monthly on what moved, why, and what we're doing about it next.
Search and social usually show inside a quarter. Programmatic, CTV, and audio need longer windows before the read means anything, both because the buying cycle they influence is longer and because the measurement is indirect. Incrementality work needs a test period we'll define with you before it starts. We'll tell you which bucket each recommendation falls into, and we won't imply they all pay off at the same speed.
A few engagements where the media moved the business.
West Monroe
B2B and finance · Strategy, web, paid media, analytics
West Monroe came to us for a website and stayed for the media program. Our team rebuilt the site around their actual positioning, then took over search engine marketing, and the paid program finished pacing to exceed its revenue target by 60 percent. Strategy, design, development, analytics, and paid media all ran in one place, which is the only reason the site and the campaigns ended up saying the same thing.
Read the West Monroe case study →A few more outcomes where we can share the number but not yet the name.
400%
Return on ad spend improvement within six months, for a direct-to-consumer eyewear brand, on a mature paid program we took over.
90%
Conversion lift for a B2B life sciences company after a site rebuild and an integrated media program.
34%
Drop in cost per acquisition in 90 days for an education publisher, after restructuring campaigns and bid targets.
116%
Increase in organic visits for a pharma B2B client, on an organic program amplified by paid.
Media behaves differently by industry, and we've spent a long time in a handful of them.
The mechanics don't care what you sell. The strategy does, and so does the definition of a conversion.
B2B and professional services
Long cycles, buying committees, and a small number of accounts that actually matter, which changes how the media gets built and how it gets read. LinkedIn does the account-targeting work without carrying the whole program, and we run Reddit and Discover for technical buyers more often than people expect, because that's where a lot of evaluation genuinely happens. Search gets built around problem-stage queries, since the buyer who already knows the category name is usually already talking to a competitor. Programmatic gets used for reach against a named target account list. The measurement is the part that decides whether any of it works: closed-won data flows back from your CRM into the platforms, so the optimization runs against pipeline instead of inquiry volume, and reporting ends at revenue instead of at cost per lead. We rebuilt West Monroe's site and then took over their search engine marketing, and a global life sciences manufacturer saw a 90 percent conversion lift after a site rebuild and an integrated media program.
Higher education
Enrollment-driven media across geographies, programs, and degree levels, where the funnel runs past the website into the admissions system and the report has to end at applications and deposits. We've been joining Slate inquiry records to behavioral data since 2013, so admissions teams can see which campaigns produced applicants, not sessions. Two of our longest-running programs sit in this category, and the work spans undergraduate, graduate, and adult and continuing education, which behave differently enough that treating them as one program is the most common mistake we inherit.
Financial services and wealth management
Compliance-aware acquisition for advisors, wealth management, insurance, and fintech, where creative review cycles, disclosure requirements, and platform restrictions on financial targeting shape the build before performance ever enters the conversation. The measurement problem here is that lead values vary so widely that cost per lead is close to useless as a target, so the programs we run optimize toward qualified opportunity value with the CRM feeding the platforms directly.
E-commerce and direct-to-consumer
Full-funnel revenue and return on ad spend across search, Shopping, social, programmatic, and retail media, where feed and catalog quality decides more of the outcome than the bidding does. We took over a mature program for a direct-to-consumer eyewear brand and improved return on ad spend by 400 percent within six months, mostly by fixing structure and measurement.
Restaurant, QSR, and franchise
Local and national campaigns tied to in-market behavior, ad fund governance across franchisees, and the geographic precision that makes multi-location media work at all. The recurring problem in franchise media is that corporate and franchisee spend compete in the same auctions, so the structure has to prevent that before the optimization can do anything useful.
Non-profit
Grant and paid programs run together, where the Google Ad Grant has to be structured around the paid account, because when the two compete for the same queries the grant quietly bids up the price of clicks you were already buying. That is the single most common way a grant account quietly wastes a paid budget. Donation and campaign pages convert quietly enough that they're the easiest thing to break, so they get protected explicitly during any restructure.
You keep the accounts, the data, and the history.
On the platforms you buy directly, everything is yours.
Google Ads, Microsoft Ads, Meta, LinkedIn, TikTok, Pinterest, Reddit, your analytics properties, your Search Console, and your tag manager all sit in accounts you own, under your billing, with your name on them. We work inside them as a user. If you leave, you keep the account, the conversion history, the audience lists, and everything the bidding algorithms have learned about your customers, which is usually worth more than any individual campaign and is exactly what agencies quietly take with them.
On programmatic, your campaigns run on our seats, and that's deliberate.
The demand-side platforms are priced for buyers at national scale, so running through our seats is what gets you premium CTV, streaming audio, addressable geo-fencing, and digital out-of-home without a platform minimum or an annual commitment. What you get in exchange for not holding the seat is visibility: line-level delivery detail, inventory reports, the site and app lists, the seller IDs, and the block lists we maintain and apply on your behalf across every account we run. We invoice working media separately from our fee, so you can see what the impressions cost and what we charged to buy them. We don't mark up media.
In your name
Google Ads · Microsoft Ads · Meta · LinkedIn · TikTok · Pinterest · Reddit · GA4 · Search Console · Tag Manager · Conversion history · Audience lists · Creative
On our seats
Yahoo DSP · Simpli.fi · Illumin
What comes with it: line-level delivery, inventory reports, site and app lists, seller IDs, block lists, full delivery history on exit.
Two other things that don't change.
You see what we see, in near real time, because reporting that only flatters the agency isn't reporting. And when the answer is that you should spend less, or pause a channel, or fix your site before adding budget, that's the answer you'll get, even when it points at a smaller engagement than the one you asked about.
Let's have a chat →Questions we get asked before a media engagement starts.
What does a paid media engagement actually include?+
Strategy and planning, the campaign build, creative production, ongoing bid and budget management, the measurement setup underneath it, and monthly reporting with analysis attached. What varies is how much of each you need. A client with a strong in-house creative team buys less of the fourth item, and a client whose tracking has never been right spends most of the first month on the fifth. We scope it against what you have, not against a standard package.
Can we hire you for an audit without committing to a retainer?+
Yes, and a lot of clients start there. The audit is a fixed fee with a defined deliverable, covering account structure, bidding, audiences, creative, landing pages, and conversion tracking, and you get a written findings document and a prioritized roadmap. It's written to be handed to whoever ends up doing the work, including your own team or another agency. We'd rather do one audit well and earn the program than sell you a program you didn't need.
Who owns the ad accounts?+
You do, on every platform you buy directly. Google Ads, Microsoft Ads, Meta, LinkedIn, TikTok, Pinterest, Reddit, your analytics properties, and your tag manager stay in your name under your billing, and we work inside them as a user. Programmatic is the exception: those campaigns run on our demand-side platform seats, which is what gives you access to premium CTV and addressable inventory without a platform minimum. You get line-level delivery data, inventory reports, site and app lists, and seller IDs on that spend, so the visibility is the same even though the seat isn't yours.
Do we need to meet a DSP minimum to run programmatic or CTV?+
No. The platforms themselves set minimums and annual commitments that are out of reach for most mid-market advertisers, and running on our seats is how you get past that. The same Netflix, Disney+, Paramount, Roku, and Hulu inventory a national advertiser buys is available to you at a budget that makes sense for your business, and the identity and addressable targeting come with it.
How do you charge?+
Management is billed as a percentage of your monthly ad spend, on a standard scale that steps down as spend grows, with a monthly minimum underneath it. The practical effect is that the rate falls as the program scales, so the bigger the budget gets the smaller our share of it becomes. Setup, audits, and one-time project work are quoted as fixed fees rather than folded into the percentage, so you know what those cost before they start. Working media is invoiced separately from our fee, which means you can always see what the impressions cost and what we charged to buy them.
We're only running search today. How do we know whether programmatic or CTV is worth it?+
Test it in a way that can actually answer the question, which usually means a geographic holdout and not a channel launched everywhere at once. We'll pick matched markets, run the new channel in some and not others, and compare total business outcomes against the new channel's self-reported conversions. That takes longer and costs a little more than a straight launch, and it's the only version of the test that tells you something you can act on. If the answer comes back no, we'll say so.
Is CTV measurable, or is it brand advertising with a better story?+
It's measurable, but not in the way search is, and any agency telling you otherwise is selling you a chart. What you can measure reliably is reach, frequency, completion, and household-level exposure, and then incremental lift in site visits, branded search, and conversions among exposed households compared with a holdout. What you can't do is give every CTV impression a last-click conversion and add it to your search numbers, which is exactly what a lot of programmatic reporting does. We report it as its own thing, with the methodology visible.
How do you handle attribution when every platform claims the same conversion?+
We separate platform-reported numbers from our own measurement and show you both. The platforms each report what they can see, which double counts by design, so we reconcile against your analytics and, where the outcome happens off-site, against your CRM. Beyond that we run multi-touch attribution where the data supports it, media mix modeling where the spend justifies the exercise, and incrementality tests where a channel's contribution is genuinely in question. The methodology and its assumptions come with the number every time, because a number without them can't be checked.
What is AI doing in our account, and are you running ads inside AI assistants?+
Two separate things, and both are worth answering. Inside your account, our own technology has been making live bid and budget decisions since 2013, reacting to auction behavior as often as every thirty minutes, and it works alongside each platform's native automation. A person still sets the targets and decides what the system is allowed to change, and in regulated categories we keep a human reviewing anything that could affect a customer negatively. Separately, we're running OpenAI Ads pixels on two client accounts, which makes us early to advertising inside AI assistants. The volume there is small and the reporting is thin, so we treat it as an instrumented test with a defined budget, not a channel we're recommending at scale.
How long before we see results?+
Search and paid social usually show inside a quarter, and often inside the first month if the problem was structural. Programmatic, CTV, and audio need longer windows before the read is trustworthy, because the buying cycle they influence is longer and the measurement is indirect. Incrementality tests need a defined test period we'll set with you before starting. Anyone showing you a straight line from month one is showing you a chart.
Will you work with our in-house team or our existing agency?+
Yes, and a good share of our media work runs that way. We've taken single channels inside a program another agency leads, run paid while an in-house team runs organic, and worked as the measurement layer underneath someone else's media buying. What we ask for is clarity about who owns which decision, because two teams optimizing the same account toward different targets is worse than either one doing it alone.
What's the minimum spend you'll work with?+
We don't publish a hard floor, because the right answer depends on the channel and the category. A search program in a category with fifteen-dollar clicks needs a different budget than a local program with two-dollar clicks, and programmatic needs enough volume for the delivery to stabilize before the numbers mean anything. Tell us the budget and the target and we'll tell you honestly whether they're compatible, including when they aren't.
What happens to our program if we leave?+
You keep every account you own, which is every platform except the programmatic seats, along with the conversion history, the audiences, and the creative. Our team documents the account structure, the bidding configuration, the tracking setup, and the exclusion lists so whoever takes it over isn't reverse-engineering it. On the programmatic side we'll provide the full delivery history and the block lists, and help you stand up your own seat or move to another buyer if that's the direction you're going.
Media work usually connects to something else.
By industry
Tell us what your media program isn't doing.
Whether that's a full program across every channel or a single audit of an account you inherited, we'd like to hear what you're seeing. We'll tell you what we'd do about it, what it would take, and whether we're the right people for it.