5 Mistakes in Facebook Advertising by Bulgarian E-commerce Businesses

Over 80% of Bulgarian online stores that “don’t see results” from Facebook ads are losing money because of the same five mistakes, not because of their budget or the algorithm. The most costly of these are optimising for the wrong goal (traffic instead of purchases) and broken tracking, which together can render an entire ad budget meaningless, even while the reports show “good” numbers.

My name is Hristian Panov—I’m a Performance Marketing consultant with over 5 years of experience managing a €2 million advertising budget for more than 90 clients on Meta and Google. Below are the five mistakes I see in nearly every audit of a Bulgarian e-commerce account, along with specific steps on how to fix them.

Mistake 1: Why is optimising for “traffic” instead of “purchases” ruining your budget?

This is the most common and most expensive mistake. A campaign optimized for Traffic or Engagement , tells Meta: “Bring me people who click.” The algorithm obediently finds cheap clicks—from people who never buy.

A concrete example from an audit of a clothing store:

• Traffic campaign: CPC €0.12, 4,000 clicks per month, 6 orders
• Same budget, same audience, optimised for Purchase: CPC €0.38, 1,260 clicks, 41 orders

Fewer clicks, nearly seven times more sales. Because Meta is no longer looking for “who clicks,” but “who buys”—the algorithm has billions of signals indicating which profiles complete a purchase.

The solution: optimise for Purchase from day one, even if you have limited data. If the account doesn’t have enough purchases (fewer than 50 per week), temporarily drop down one level—Add to Cart or Initiate Checkout—but never Traffic for e-commerce. The objective determines who Meta sends to you.

Mistake 2: What do you lose with broken or missing Pixel/CAPI tracking?

If you’re optimizing for purchases but tracking is broken, Meta optimizes blindly—bidding on fake or incomplete signals with your money. This is the silent mistake: the reports look normal, but 20–30% of the actual data is missing.

Three common issues I find:

1. Pixel only, without the Conversions API (CAPI). Browser blockers and iOS restrictions eat up 20–35% of the events that Pixel alone fails to send. CAPI sends them server-side, bypassing the blockers.
2. Event duplication. Pixel and CAPI send the same purchase without a shared event_id Meta counts it twice. I’ve seen accounts with ~30% inflated conversions because the event_id doesn’t match between the browser and the server.
3. Missing parameters. A purchase without a value or currency → Meta cannot optimize for ROAS because it doesn’t know how much each sale is worth.

The solution: Pixel + CAPI via a server-side setup (Stape + GTM), with a single event_id, sent to both sources, plus an External ID as a second layer for matching. Result: deduplication drops below 5%, and Meta gets the full picture. This is a one-time setup costing ≈ €420, which often pays for itself within the first month.

Mistake 3: Why do “professional” creatives often perform worse?

Bulgarian e-commerce businesses often confuse “beautiful” with “selling.” A polished product shot on a white background, a large logo, a corporate tone—it looks good in a portfolio, but performs poorly in the feed.

What works in real-world data:

• Native content — a video shot on a phone that looks like an organic post, not an ad. It often beats the studio-produced product by 2–3 times in CTR.
• The first 3 seconds — if the video doesn’t stop the scroll immediately, the rest doesn’t matter. Hook before branding.
• UGC (user-generated content) — a real person holding and using the product converts better than any render.
• Text on the image — a specific benefit or situation, not “high quality” or “best price.”

A mistake I see all the time: a single creative run for months on end until it dies of creative fatigue (frequency rises, CTR drops).

The solution: a library of up to 6 communication approaches within the framework of objective advertising needs — different angles (problem→solution, desired result, social proof, curiosity), not six versions of the same thing. Test them, keep the winners, replace the tired ones.

Mistake 4: Why does scaling up too early kill a profitable campaign?

You discover a profitable campaign running at €20/day, get excited, and scale it up to €100/day overnight. The next day, ROAS plummets. This isn’t bad luck—it’s a predictable mistake.

Meta has a learning phase. A sudden change in budget (over ~20–30%) resets the learning process—the algorithm essentially starts over, resulting in unstable and expensive performance. Additionally, the audience gets exhausted: the same budget targeted at a segment that’s too narrow drives up frequency and CPM.

A specific scaling pace that works:

• Increase the budget by 20–30% every 3–4 daysnot all at once.
• Monitor frequency—over 2.5–3 within a week for a cold audience is a sign that you’ve pushed too hard.
• For a bigger jump, duplicate the winning campaign into a new one with a broader audience, rather than forcing the budget on the existing one.

The solution: scale gradually and judge by POAS, not daily ROAS. A stable campaign at €60/day beats an unstable one at €100/day that constantly restarts.

Mistake 5: Why shouldn’t you blindly trust the Opportunity Score?

Meta displays an “Opportunity Score”—a number from 0 to 100 with recommendations on how to “improve” your account. The problem: these recommendations optimise for Meta’s goals (higher spend, broader targeting, Advantage+ everywhere), not necessarily for your profit.

Typical recommendations that often do more harm than good:

• “Remove targeting”—Meta wants the freedom to spend broadly; for a niche product, this blurs the audience.
• “Increase the budget”—recommended regardless of whether the campaign is profitable.
• “Merge ad sets”—sometimes sensible, but not because the score drops by 4 points.

I’ve seen businesses chasing a 100/100 score and, in the process, destroying a structure that worked. The score is a tool for attention, not a verdict.

The solution: treat the Opportunity Score as a list of hypotheses to test, not as commands. Every recommendation should pass one test: “Does this improve my POAS, or Meta’s spend?” Implement only what passes the test—and always against your own data.

What is the true cost of these mistakes taken together?

A realistic estimate for a store with a €1,500/month budget that makes all five mistakes:

• Optimisation for traffic: loses ~60% of potential orders
• Broken tracking: Meta optimises blindly, another ~25% loss in efficiency
• Fatigued creatives: CTR drops, CPM rises
• Rash scaling: periodic ROAS crashes
• Blindly following Opportunity Score: a broken working structure

The same budget, with the same product, after fixing the five issues—two to three times more orders. Not because more was spent, but because every euro was targeted correctly.

FAQ — Frequently Asked Questions About Facebook Advertising for E-commerce Businesses

1. How many purchases per week are needed to optimize for Purchase? Ideally, 50+ purchases per week at the ad set level to exit the learning phase. Below this threshold, temporarily optimise for Add to Cart or Initiate Checkout, but not for traffic.
2. Do I really need CAPI, or is Pixel enough? CAPI isn’t an option—it’s a necessity for e-commerce. Pixel alone loses 20–35% of events due to ad blockers and iOS restrictions. Without server-side tracking, Meta optimises with incomplete data.
3. How often should I change creatives? When frequency increases and CTR drops—that’s creative fatigue. In practice, every 2–4 weeks for active campaigns. Maintain a library of different angles, not variations of the same one.
4. How quickly can I scale a profitable campaign? A 20–30% budget increase every 3–4 days. More aggressive increases restart the learning phase and destabilise results. For greater growth—duplicate the campaign, don’t force the budget.
5. Should I follow the Opportunity Score recommendations? Treat them as hypotheses to test, not as commands. They often optimise for Meta’s spend, not your profit. Implement only what improves POAS.
6. Why are my ads generating clicks but not sales? It’s almost always one of three things: optimisation for the wrong goal (traffic instead of purchases), broken tracking, or a mismatch between the ad and the landing page. Clicks without sales are a symptom of incorrect setup, not a “bad audience.”