I'm Dimitri Bukys, a senior performance marketing specialist running Google Ads and Meta for ecommerce and DTC brands. Most of the work happens before I spend anything, because a ROAS target built on the wrong margins is a comfortable way to lose money. I want your costs, your returns and your real breakeven first. The ads come after that, and by then they have a number to clear.
Dimitri Bukys
8years in paid acquisition
3.7×blended ROAS, DTC account, year two
−40%spend, revenue held within 5%, backtested on 402 ads
Three years on one account, through the point where every channel they relied on stopped working at once.
One of these is probably you right now.
01
Spend goes up, profit doesn't.
Every increase in budget drags the blended number back down, so the account sits at a ceiling nobody chose. Scaling revenue that is structurally unprofitable is a worse outcome than not scaling at all, and it usually takes a quarter to notice.
02
Shopify and the ad platforms tell different stories.
Meta claims the sale, Google claims the same sale, and between them they claim more revenue than the business actually banked. Budget then gets decided by whichever dashboard happened to be open.
03
The account has history, not a system.
Years of activity and very little reusable learning. Tests get launched without a hypothesis, without a rule for killing them and without a written conclusion, so nothing that was learned last quarter is available this one.
04
Creative only appears after performance drops.
Nothing is queued before the winners burn out, and the next brief is usually whatever came out of the last status call rather than what the data said was working.
05
You suspect the ads are not the bottleneck.
It can sit in the offer, the margin, the product page or the returns rate. The difficulty is that almost everyone you could ask about it is paid to tell you the answer is more ads.
06
Nobody in the building understands the account.
The last partner held the knowledge and left with it, and the handover was a login plus a dashboard nobody can interpret. Growth is now hostage to whoever happens to be running the campaigns.
Different starting points, same first move: work out what the business can genuinely afford to pay for a customer before touching a single campaign.
How I work
Most of this happens before a campaign goes live, and that is deliberate. Anyone can turn ads on. What decides whether they make you money is the work underneath: what you sell, who buys it, how a sale actually completes, and what a customer is genuinely worth to you once discounts and returns are taken out.
The business first
I go through the offer, the buyers, and whatever is already bringing work in. Then I audit the ad accounts and analytics against that reality, not against what the account assumes is happening.
You know what is actually blocking growth, so the budget goes at the real problem first.
Make the numbers agree
If campaigns cannot be connected to real revenue or bookings, that gets sorted first. Sometimes the platform setup is enough. When it is not, I build the connection with APIs and whatever code the job needs.
Meta and Google will both claim the same sale, and their totals together will never match what your bank did. I reconcile against your real revenue and tell you which number I am steering on.
You make budget decisions from one version of the truth instead of whichever dashboard was open.
Set the floor
Real margins, AOV, returns, and what a conversion is genuinely worth give us the breakeven ROAS or maximum acquisition cost. Everything after that gets measured against the business's own number.
You know the exact point where buying another customer stops making money.
Find where growth could come from
I look for waste inside the account and openings outside it. Reviews show what buyers care about, forums show the words they use, and competitors show what the category keeps getting wrong.
You get campaign angles grounded in what buyers already care about, not ideas pulled out of the air.
Build a weekly engine
A good account needs a steady stream of new creative, fresh angles, and experiments. Every test starts with a reason and a number to beat, and win or lose it leaves a reusable answer.
The business gets a repeatable pipeline of better bets instead of waiting for one lucky campaign.
Scale without losing the plot
Whatever misses the floor gets cut. Whatever clears it earns more budget. Results feed into the next research round, so the account keeps learning instead of becoming a museum of old campaigns.
Budget moves toward profitable demand while weak spend gets cut before it compounds.
Nearly everyone in this market gives the audit away and uses it to sell the retainer, which makes it a sales call with a document attached. Mine is paid, and that is the only thing that lets it reach the conclusion that you should not hire me.
What I need from you
Read access to Google Ads, Meta, GA4 and your store analytics
Your real numbers: cost of goods, shipping, discounting, returns and average order value
One hour to walk me through how a sale actually completes
What you get back
Your true breakeven ROAS and maximum acquisition cost, built from your margins instead of a platform default
What the platforms claim, reconciled against what the business actually banked
A diagnosis of the account, the tracking and the path after the click
A prioritised 30 to 60 day plan, in the order I would work it
How long it takes
Five working days from the moment access lands, then one call to walk you through it. You keep the document whichever way the decision goes.
What it costs
$500
Fixed. It comes off your first invoice if we go on to work together. And if the honest answer is that ads are not your bottleneck, the document says exactly that, which is a real result rather than a failed sale.
There was no paid acquisition when I arrived, only campaigns the owner had set up himself with no way of telling what they earned. I built the structure, the tracking, the research behind the messaging, and the testing framework the account still runs on.
Google took six months to crack, and by the autumn the account had dropped below breakeven. Rather than spend another quarter experimenting with the client's money I bought outside counsel, made the changes, and November, the weakest month of the season, came back strong.
This season sold the stock out, then ran a month with no ads at all because there was nothing left to sell. Breakeven ROAS has since been rebuilt from 2 to 3 and the target to 4, because the real cost of discounts and returns turned out higher than anyone had allowed for.
From bookings nobody could trace to a system that could be steered.
No booking could be traced back to the campaign that caused it, so every scaling decision was a guess. I went through the analytics, the tag manager and the booking engine, found what was breaking, rebuilt the events and fed correct data back into Google Ads. Revenue moved immediately and November became the best month the business had ever had.
The caveat, because it belongs here: I then proposed separating branded from non-branded search, the rebuild destabilised the account, and repairing it took two to three months. The structural call was right and the sequencing was wrong, which is the sort of thing worth knowing about someone before you hire them.
Rather than guess at replacements I went back through their own CRM history to find where customers had actually come from over the years. That surfaced a category of marketplace nobody in the market was buying advertising on, so I tested all of them, found the one that produced real customers, and bought every placement it had.
Then the customers were split by occasion, and the useful discovery was lead time. Corporate bookings for New Year are decided two to three months ahead, so the offer has to be live by the end of August. Each occasion got its own window and its own offer, and the gain came from rebuilding what was offered rather than redesigning the page.
The person who sold you the account is rarely the person running it, and when the team gets reassigned the knowledge walks out with them. Their model also needs your budget to grow whether or not growing it is the right call this quarter.
Instead of a junior or a cheap freelancer
Cheaper by the hour and a good deal more expensive by the quarter. The learning happens on your budget, and the mistakes that matter are the ones that take a whole season to become visible.
Instead of a full-time hire
$90k to $120k a year before recruitment costs, and one person's blind spots quietly become your entire strategy. You also need to already know this job well enough to manage someone doing it.
What you get here
Me, on your account, five clients at a time and no more. I also take on nobody new in the quarter we start, so the months that decide whether this works are not the months I am learning someone else's business. Direct access with nobody in between, and everything I build written down so it stays yours the day we stop working together.
What it costs
Published, so you can decide whether this is worth a call before booking one.
01 / Diagnose
Paid diagnosis
The account, the tracking, the funnel, the offer and the economics, ending in a prioritised plan and a straight recommendation: scale, repair or stop.
For anyone who wants a decision before a commitment.
Built where there is nothing, rebuilt where there is a mess, then run: campaign structure, measurement, research, a weekly testing rhythm, creative direction, and budget decisions against your real floor.
Does not include store development, email or lifecycle, or creative production beyond direction and static assets.
From $4,500first three months · $1,500/mo minimum after
One table, because the fee is not one number. What you pay depends on what the account returns.
Monthly ad spend
Below your breakeven ROAS
Between breakeven and target
At or above target
Up to $50,000
$750half the minimum
$750 plus a share of the restby the formula below
$1,500, or 15% of spend if that is greater
Above $50,000
Half the agreed minimum
Half the minimum plus a share of the rest
Negotiated on the same rule, with the percentage stepping down as spend rises
Below your breakeven I am on half pay.
That is the part worth reading twice. If the account is not yet making you money, neither am I, and I carry that until it does.
The middle column, in fullfee = half the minimum + (the rest × (achieved ROAS − breakeven ROAS) ÷ (target ROAS − breakeven ROAS))Cover half the distance from breakeven to target, earn half of what sits above the floor. At breakeven it pays the floor and nothing more; at target it pays all of it.
Breakeven and target both get written down before we start, along with where they are measured and over what window. So I do not reach my own minimum until you are past breakeven, and not the full fee until you hit the number we agreed.
The accounts, the campaigns, the creative and the reports belong to you and stay with you. Either of us can end this with thirty days notice, and inside those thirty days you get a written handover for whoever comes next, the full change log, the reporting logic with the calculations shown rather than hidden in a tool of mine, and every test with what it showed.
An account that only works while I am in it is a hostage arrangement, not a service.
You are already spending, ideally more than $10,000 a month, so there is room to run real experiments
You can give me margins, returns and discounting, not just ad-platform access
Somebody can produce creative, or there is budget to
This works badly when
The budget is small enough that a single test uses up the month
The real numbers are off limits, which makes any ROAS target a guess
You want a pair of hands to execute a plan that is already decided
You need daily presence in a chat rather than decisions and results
I turn down
Gambling, and the grey categories sitting next to it
Anyone who wants a guaranteed result written into the contract, because the honest version of that promise needs control over your pricing, your margins and your stock
What I don't do
Two kinds of no, and it is worth knowing which is which.
A deliberate boundary
Amazon and retail media, SEO and content production, sourcing and managing influencers, running your shop, inventory and purchasing decisions, building sites and landing pages. I diagnose the page and specify the fixes in writing, then someone who owns the site implements them.
Not offered yet, and this line changes when it does
Email, SMS and lifecycle. Server-side tagging and the Conversions API, which I will diagnose and specify but will not claim to ship. Offline conversion imports.
TikTok and LinkedIn I have tried, and the results did not justify recommending them, so I will not sell you a channel I do not rate.
Why is the diagnosis paid when everyone else gives it away?
Because a free audit is a sales call with a document attached, and it can only ever end in one recommendation. Charging for it is what allows it to conclude that your ads are fine and your problem is somewhere else. You are buying a decision, not a proposal.
What if the ads turn out not to be the problem?
Then the document says so, in writing, and tells you where the bottleneck actually sits. That happens often enough that it is worth planning for. It is also why I want the margins and the returns data before I look at a single campaign.
Do you really need my margins?
Yes, and it is the one thing I will not proceed without. Without cost of goods, discounting and returns there is no breakeven ROAS, and without a breakeven ROAS every target on the account is somebody's guess dressed up as a number. If sharing that is not possible, I am the wrong person and I would rather say so on the first call.
Who makes the creative?
I do the research, the angles, the briefs and the static production, and I can produce a great deal of it quickly. Video I direct rather than shoot: I will tell your team or your editor exactly what a winning video needs to contain, based on what has already worked, but I am not a production studio. If you have no way to produce video at all, say so early, because it changes what the plan can rely on.
Do you implement the tracking yourself?
Yes. Tag manager, events, cross-domain, the analytics side and the reconciliation back to real revenue, including custom work through APIs where the platforms are not enough. The exception is changes that have to be made inside your site or theme code, where I specify precisely what is needed and your developer ships it.
Who owns the accounts and the data?
You do, always. Everything is built inside your own ad accounts, your own analytics and your own tag manager, and the documentation goes with it. If we stop, you keep a working system and a written explanation of how it works rather than a login and a shrug.
How often will I hear from you, and how fast?
Reporting is weekly and it explains decisions rather than listing metrics. I work to 16:00 in Almaty, which is midday in London, and inside those hours a message usually gets an answer within the hour. After that it gets answered, just not quickly. Calls happen in any timezone by appointment, evenings included, and high-spend periods get extended cover agreed in advance.
Do you guarantee a ROAS?
No, and I would be careful with anyone who does, because the honest version of that promise requires control over your pricing, your margins and your stock. What I do instead is put my own fee behind it: below your breakeven I am on half my minimum, and I do not earn the full fee until you hit the target we agreed.
Can you work alongside my agency?
Yes, and I also work white label for agencies directly, so the situation is familiar from both sides. What matters is that one person owns the spend decision. Two parties optimising the same account against different targets is how budget quietly goes missing.
What happens after the diagnosis?
You take the plan and run it yourself, or you take it to somebody else, or we carry on together and the $500 comes off your first invoice. All three are fine. The diagnosis is built to be useful on its own, which is the only way it can be honest.
The longer version
The short version is seven years of paid ads. The longer one has a war in it, so here's the longer one.
Where I learned the craft
I started in-house at Paper Planes in Moscow, working through the agency on campaigns for Pepsi, Coca-Cola, McDonald's, Bosch and major retailers, managing more than $1M in ad budget between them. Then I found surfing, the office stopped making sense, and I went remote for good.
In December 2020 I landed Casa Picassa, a Moscow events venue and the first client that was fully mine. It began alongside Paper Planes and became the longest relationship on this page.
Paper Planes continuesCasa Picassa begins
The year everything moved
I'm Ukrainian, born and raised in Russia. When the war started I went out to protest, came close to being arrested, and left the country for good. Paper Planes and I parted over opposite political positions, while Casa Picassa stayed with me through it.
Paid traffic dried up, so I stopped thinking like a media buyer and became the whole marketing team, finding customers through channels most people never look at.
24 Feb 2022 · Russia attacks UkraineLeft RussiaPaper Planes endsCasa continuesSupport Ukraine
One line ends. Another begins.
In March 2025 Casa Picassa and I parted on good terms. A month later I started with Globaltic in Manchester, and the international work properly began.
Casa Picassa endsGlobaltic beginsWavehouse begins
The journey continues
I've never wanted to run this the same way two years running. Lately that means building AI deep into the research and testing so grunt work stops eating hours and the judgment gets sharper. The work I'll be proudest of still feels ahead of me.
Surfing, skating, mountains, too much hip hop. Certified lifesaver of people and brands since 2019.
Tell me what you're working with.
Send over the business, what you're spending a month, and what you actually want paid acquisition to do next. I'll tell you honestly whether I can help.
