Diagnosing a Growth Ceiling: What Broke (and Fixed) a Premium DTC Brand's Best Campaign

A client came to us with a familiar-sounding problem: a campaign that used to convert reliably had quietly stopped working. Not a crash — a slow bleed. Spend was flat, clicks were flat, and purchases had fallen off a cliff. The instinct in most agencies is to blame the creative. It wasn't the creative.

This is a case study from one client account we manage — a premium DTC and B2B brand selling handcrafted, realistic artificial trees to hospitality venues, offices, and residential customers. Because the account is under NDA, we're not naming the brand. We built this account's paid acquisition system from the ground up; this case zooms into one specific diagnostic episode from that broader work, not the full build. Everything below — the numbers, the campaign structure, the diagnosis — is real.

The business: a high-ticket product that's hard to evaluate online

The brand sells handcrafted artificial trees ranging from a few hundred dollars to well over $5,000 for large custom installations. Buyers fall into two very different groups: individual homeowners furnishing a living room, and businesses — restaurants, hotels, spas, offices, event venues — outfitting a commercial space. Same product, completely different purchase logic.

That combination — high price point, low purchase frequency, a product that's genuinely hard to evaluate from a product photo, and two buyer types with opposite priorities — is exactly the kind of account where "just run more ads" stops being useful advice. When something breaks, it's rarely one thing.

Why "the ads stopped working" is almost never the real diagnosis

When a previously scalable acquisition system stops scaling, we don't start with creative. We start by walking the full chain: traffic quality, landing page and PDP experience, offer clarity, trust signals, checkout friction, attribution accuracy, and — often overlooked — whether the audience itself quietly changed underneath the campaign.

For a product like this, three diagnostic layers matter more than most accounts:

• The visualization gap. A buyer isn't evaluating "a tree" — they're trying to picture it in their specific restaurant, lobby, or living room. A generic product shot doesn't answer that question, and no amount of ad spend fixes a page that can't.
• The commodity trap. "Artificial olive tree, 8 ft" is a search query, not a brand. Without a clear reason to buy from this business specifically — craftsmanship, custom sizing, business support, warranty — the buyer just price-shops.
• The B2B path that doesn't exist. Someone who needs 15 trees for a hotel lobby isn't going to "Add to Cart." If the only conversion path on the site is a single-unit checkout, that demand quietly disappears before it's ever counted.

All three of those are real, ongoing work on this account. But the episode below — the one with clean before/after numbers — came from somewhere simpler and easier to miss.

The flagship diagnostic: an 83% revenue collapse hiding in a routine geo update

The account's best-performing catalog sales campaign was running across a mixed international audience — primarily US, with meaningful volume from Australia and the UK. In late May, as part of a routine scope adjustment, the campaign's geo targeting was narrowed to US-only.

On paper, that's a small, boring change. Here's what it actually did to the campaign, week by week:

Top-of-funnel volume barely moved — Add to Cart stayed in the same range before and after. The break was entirely in the bottom of the funnel: checkout-to-purchase collapsed from 19–42% down to 4–14%. At the campaign level, monthly revenue went from $91,165 in May to $15,448 in June — a drop of roughly 83%, on essentially the same ad spend.

That pattern is the tell. If a creative or fatigue problem were the cause, Add to Cart would have dropped too — people wouldn't have engaged in the first place. Instead, people were engaging at the same rate and then failing to convert at checkout. That's an audience-quality problem, not a creative problem: the AU/UK segment being removed carried a dramatically higher average order value ($4.8K–$8K) than the US-only traffic that remained. Cutting it didn't just remove volume — it removed the buyers who were actually completing high-ticket purchases.

The fix — and why we didn't just revert it

The obvious fix is "put AU/UK back." We didn't do that immediately, because the goal wasn't to return to the exact prior state — it was to understand which part of the audience mix was load-bearing and rebuild deliberately. The campaign was updated to reintroduce international reach in stages: Canada, Germany, and Switzerland were added back first, with Australia and UK — the two highest-AOV markets — held out for a separate, more controlled test rather than reintroduced all at once.

The result: not fully back to the old audience mix, but already outperforming it on cost

Here's the honest part of this case: purchase rate and AOV have not returned to the original 19–42% / $4.8K–$8K levels yet, because the audience mix genuinely hasn't been restored to what it was — AU and UK are still out. Judged purely on purchase rate, the campaign today would look unfinished.

But purchase rate isn't the metric that actually matters here — cost per acquisition is, and that's where the diagnosis paid off:

Current CPA is not just recovered from the June collapse — it's now lower than the original May benchmark, on a narrower audience that still excludes the two highest-value markets. That's a meaningfully stronger acquisition engine than the one that existed before anything broke, even though the headline purchase-rate number hasn't caught up yet. Reporting only PurRate here would tell a worse story than what's actually true; reporting only CPA would hide that the audience rebuild isn't finished. Both numbers, together, are the honest picture.

What this episode reveals about diagnosing paid media problems

The lesson isn't "don't change your targeting." It's that a change which looks purely operational — narrowing a geo, simplifying a setup — can silently remove the specific segment of buyers an account depends on, and the only way to catch it is checking funnel-stage metrics against each other, not just the topline number. Add to Cart told us the ads still worked. Purchase rate told us the audience had changed. Revenue told us how much it mattered. None of those three numbers alone would have pointed to the actual cause.

This is the same lens we apply across the rest of this account — the commodity trap, the visualization gap, the missing B2B conversion path. None of it starts from "what should the ad say." It starts from asking which specific number moved, and why, before touching a single creative.

A quick diagnostic checklist for your own account

If a campaign that used to convert has quietly stopped, before you touch the creative, run these four checks:

1. Plot Add to Cart and Purchase Rate on the same timeline. If Add to Cart is flat or growing while Purchase Rate is falling, the problem is downstream of the ad, not the ad itself — look at audience quality, checkout, or the offer, not the creative.
2. List every targeting change from the weeks before the drop. Geo edits, interest changes, lookalike source swaps, budget reallocations — even ones that felt minor or purely operational at the time.
3. Break revenue down by segment before and after the change. A flat topline number can hide a specific high-AOV segment disappearing entirely, especially in accounts with a wide price range.
4. If you're mid-recovery, judge by CPA, not just Purchase Rate. Purchase Rate stays depressed as long as the audience mix is incomplete, even once the underlying account is healthy again. CPA tells you whether it's already profitable to keep spending.

Want this diagnosed on your own account?

If a campaign that used to work has quietly stopped converting, the fix is rarely "try new creative." It's usually one specific number, in one specific place in the funnel, that changed underneath you. Get a free marketing audit from SciGrowth and we'll walk your account the same way — funnel stage by funnel stage — before touching a single ad.

FAQ

How do you tell an audience problem apart from a creative problem?

Compare top-of-funnel engagement (clicks, Add to Cart) against bottom-of-funnel conversion (checkout-to-purchase) over the same period. A creative or fatigue problem shows up as declining engagement across the whole funnel. An audience-quality problem shows up as stable or even growing top-of-funnel activity with conversion collapsing specifically at checkout — because people are engaging, but they're the wrong people.

Why not just revert the geo change immediately?

Reverting tells you the symptom is fixed, not why it happened or which part of the audience was actually load-bearing. Rebuilding in stages — adding markets back one at a time and measuring each — turns a one-time recovery into a permanent diagnostic: you end up knowing exactly which segments drive AOV and which just drive volume, instead of guessing.

Why report CPA instead of purchase rate as the recovery metric here?

Because they answer different questions. Purchase rate tells you how efficiently a given audience converts; it will stay depressed as long as the audience mix is incomplete, even if the campaign is healthy. CPA tells you what it actually costs the business to acquire a customer — which is the number that determines whether the campaign is profitable right now, independent of whether the audience rebuild is finished.