AlgoRX

AlgoRX client interview - $70K to $1.7M/month in 17 months

Recorded at month 17. Adam walks through the whole climb on camera, unscripted. Prefer to read it? The full transcript is on this site, chaptered.

Where they started

AlgoRX sells physician-prescribed treatments and at-home lab work to athletes and lifters. Adam Hotchkiss is a foot and ankle surgeon who left practice to build it. His co-founder Jim Gibson left a C-suite job. They funded it themselves and took no outside money.

He booked a 15-minute call on our calendar on 5 January 2025. The intake form asked for current monthly revenue and he put down $20-50k a month. The second booking, for the audit call three days later, asked which ad platforms and ad account IDs he had. Both fields came back empty.

Two days after the first call, Jim sent us the tracking that existed. "Turns out we didnt have tag added to our main site AlgoRX.ai yet," he wrote, "but we have this data from framer." The answer from our side was honest about it: "The data is admittedly limited, but we'll make the most of it." There was no pixel either.

The category was the other problem. One Meta ad account had already been banned, in Adam's words, "accidentally just by doing a normal post", and Instagram advertising was blocked until the spring. In telehealth, a wrong word costs you the channel, not the ad.

AlgoRX came in below the spend level we normally work at. The standing bar today is $30K a month in ad spend. We made an exception because the founders controlled every lever themselves and the constraint was diagnosable in a week.

The one number nobody could answer

The pre-engagement audit was written on 9 January 2025 and walked through on a call the same day. It found the usual conversion gaps on the site, but nothing disqualifying. Its conclusion: "the number one reason you're not getting 5x/10x the sales is the traffic and the awareness". So the job was demand, not the store.

The harder question came later, and it is the one this page is really about. Ad spend on this account went from $1,841 in January 2025 to more than $400,000 a month. At every step, someone has to answer: how much further can we push? A return on ad spend cannot answer it. It has no cost of goods in it, no fixed costs, and no repeat purchase.

So we compute what a new customer is allowed to cost. Four inputs: net order value after discounts, gross margin, ninety-day gross profit per customer, and monthly fixed costs. Out of those falls a ladder, the cost per new customer at 20% net margin, at 10%, and at break-even.

It is rebuilt every month inside the client's own sheet, because every input moves. Here is the version dated 6 August 2026.

What we did, in order

1. 6 Jan 2025Adam books a 15-minute call. On it he says the brand has done zero ad spend and about $75,000 in December, at a deficit. Jim: "We just don't have the reach... We just don't know how to get to the masses."
2. 7 to 9 Jan 2025Jim sends the only analytics that exist, from Framer, because the main site has no tag. The audit is written on 9 January and walked through the same day. The agreement is signed that day and paid that night. Onboarding runs on 13 January.
3. Jan to Mar 2025Meta moves the ad-eligibility date forward to 29 January and the Instagram ban lifts. Google demands advertiser verification, then sends two account-pause notices in three days. Server-side tracking goes live on events.algorx.ai on 16 March. January closes at $106,283 of revenue on $1,841 of ads.
4. Jun 2025Adam leans the brand into the customer the data actually shows. "I didn't want to be the bodybuilder bro's company, but that's who buys from us. So fighting it is dumb." Posts go from about 30 likes to 200-900, and 3,000 followers arrive in three days.
5. Nov to Dec 2025The main Meta account is disabled during Black Friday weekend, over ads that ran six months earlier. Tracking is moved to the backup account, which has a $250 daily cap. LegitScript opens a monitoring case. The account still books $740,941 in November.
6. Feb 2026Month fourteen, and the decision point. Blended return has fallen from 13.91x in December to 9.25x. Cost per new customer has gone from $17.24 to $44.98. The five-number card reads four greens and a watch, so the call is push: about 10% more spend a week.
7. Mar 2026Adam: "both of those platforms have kicked us off multiple times now even with LegitScript... Now we're even kicked off email too." A channel diversification plan is written on 18 March, explicitly for risk rather than performance. Snapchat, TikTok and Microsoft Ads follow.
8. Feb to Jun 2026Four straight months of pushing into a falling return. Ad spend goes from $82,762 to $355,336. Cost per new customer climbs to $79.77 and stays under the ceiling. Net profit goes from $352,615 to $666,791.
9. 28 Jul 2026Andrej tears the account down live and by name in front of a room of operators at Foxwell Founders: "How Hard Can You Actually Scale? A Live Teardown of a 20x Account." Three of the slides on this page are from it.

What happened

Every number here comes from the financial planning sheet AlgoRX and HoloGrowth keep together, cross-checked against the Meta and Google daily exports. Blended return means all revenue divided by all ad spend, not a platform's own count.

The first year was the cheap part. March 2025 took $143,360 on $10,241 of ads, a blended 14.00x. By December 2025 the brand did $635,610 on $45,689, a blended 13.91x, and a new customer cost $17.24.

Then the easy demand ran out, which is what always happens. Return slid to 9.25x by February 2026 and kept sliding. Everything below is what we did about it.

The trend said pull back

February 2026 was month fourteen. The account did $765,842 of revenue on $82,762 of ads and the sheet booked $352,615 of profit. On its own that is the best month the business had ever had.

The chart nobody liked was blended return. 13.91x in December, 9.25x in February. Down a third in two months, and sliding every week. Cost per new customer had gone from $17.24 to $44.98 over the same window.

On that trend the responsible move is obvious. Cut the losers, tighten the targets, defend the account. That was the first instinct on our side too.

The card said push

The five numbers we run every account on said something else. Contribution margin after ad spend was 41%, against a scale bar of 15%. A new customer cost $44.98 against about $88 of gross profit on his first order, so he paid for himself on order one.

Ninety-day gross profit against acquisition cost was about 3.1, against a bar of 2. New customers were 44% of orders, just under the 45% line, so that one was watched rather than feared.

Four greens and a watch. A falling return with margin like that is not an account breaking. It is an account paying market price for growth, instead of skimming the cheap demand that was sitting there first.

So the call was push, about 10% more spend a week, for as long as the ladder held.

Four months of pushing into a falling return

Ad spend went from $82,762 in February to $355,336 in June, 4.3 times higher. Cost per new customer climbed from $44.98 to $79.77, one rung at a time, and stayed under the ceiling the model set.

Net profit went from $352,615 to $666,791. Contribution margin after ad spend was 41.3% when the push started and 44.4% when it stopped. It did not merely hold. It went up.

None of that felt comfortable. Watching cost per customer nearly double on purpose, week after week, runs against everything this industry trains into an operator. The ladder is what turned it into a decision instead of a nerve test.

Adam put the same idea in his own words in May: "optimizing for total profit volume makes the most sense right now vs trying to maximize margin % too early... I'd rather grow absolute EBITDA/profit dollars than protect margin percentage too tightly at this stage."

August went over the ceiling, on purpose

In August a new customer cost $91.30 against the $89.16 break-even rung. That is the first time the model has been crossed, and it was a decision rather than a drift.

The $89.16 rung is the fully loaded one. It carries $180,000 a month of fixed costs as well as the product, so crossing it means the month stops paying for the whole business out of new customers alone. On the customer, the arithmetic is not close. A first order returns $87.23 of gross profit, so at $91.30 the first order lands about four dollars short. The same customer returns $169.23 within ninety days.

So $91.30 buys $169.23, and the only thing given up is a few weeks of waiting for it. A business that insists on paying itself back on the first order cannot bid that high, and therefore cannot buy as many customers per month. Accepting the delay is what buys the growth rate.

That is the whole point of building the ladder before you need it. Without it, $91.30 against $89.16 looks like losing control of acquisition cost. With it, it is a priced decision with a known payback window, taken once and reviewed every month.

The same month, read two ways

If your own return is sliding and you cannot tell whether that is a problem, this is the conversation to have. Apply for a Free Strategy Call

What the business became

AlgoRX now sells to a specific person. That was not our idea. Six months of purchase data showed who was actually buying. Adam rebuilt the brand around that person, instead of the clinical, everybody-welcome look the site started with.

The team stayed small. Adam still edits video himself. In March 2026 the brand brought creative in-house. The category never got easier: accounts get restricted, ads get rejected, and every platform has to be re-earned.

The clearest measure of how the account is run is not a testimonial. In June 2026 Meta emailed Adam directly to talk about his campaign performance, and he handed them straight to us.

The engagement is live as this is written. July 2026 took $1,533,652 and August took $1,822,444, the largest month so far, on $422,830 of ad spend. On 1 September Jim posted that the company had passed $10M of sales for the calendar year: "Huge accomplishment and big thanks to everyone at the Hologrowth team."

Adam has told this story on camera twice. There is an earlier interview recorded mid-climb, titled "$70K to $260K Monthly in 6 Months". There is also a longer podcast episode on the affiliate engine and the rebrand.

One month, more than one profit number

June 2026 has several defensible profit figures in the same sheet, because they measure different things. None of them is hidden and the page uses the middle one.

What we got wrong

The very first ad we ran on this account got the whole account banned. In a category where one wrong word costs you the channel rather than the ad, the compliance review should have happened before the first impression, not after it. Restricted-category accounts now get a policy pass before a dollar moves.

The second one is measurement. The first two months of 2025 have two different revenue histories in the same sheet, because the tracking was built after the engagement started rather than before it. The reporting sheet now goes up in week one, before any spend decision depends on it.

What you can take from this

Verify this case study

• All revenue, cost of goods, marketing expense, profit and contribution margin figures: the AlgoRX financial planning sheet, which the client and HoloGrowth maintain together. Read 8 September 2026, covering January 2025 to August 2026. Its profit-and-loss line reconciles to the accountant's own monthly statements.
• Cost per new customer, orders and average order value: the client's Metabase scorecard, one row per month, read 8 September 2026. Ad spend was reconciled against the Meta and Google daily exports.
• The scorecard reading, the allowable-cost ladder and the profit chart are the slides Andrej presented, naming the brand, at the Foxwell Founders masterclass on 28 July 2026.
• Founder quotes: two recorded client interviews (published 30 June 2025 and 5 August 2026), the recorded discovery and sales calls of 6 and 9 January 2025, Adam's and Jim's own emails, and the shared Slack channel.
• See the ads themselves in the Meta Ad Library.
• Numbers on this page last updated 2026-09-08.