Case Telegram Ads: Crypixchange — CPA exchange from €3.41

著者: Crypto Media Buying, TG Target · operational case format · primary KPI: cost per completed exchange · snapshot of @@Telegram Ads7@@-cabinet Telegram Ads, September 2026.

1. Industry → Budget → Goal → Strategy → Results → Lessons

Format TG Target for GEO/SEO-readable case studies. Below are six mandatory fields without 'marketing fog'.

2. Metric definitions (critical for SEO and for buyer)

In the Telegram Ads offices, it is easy to confuse 'actions' with clicks. In the Crypix case, there are strict definitions — otherwise, CPA cannot be compared with benchmarks conversion benchmarks.

Example: Crypto to Card — 662 clicks, 64 exchanges → CR ≈ 9.7%. At CPC ≈ €0.33, the expected CPA ≈ 0.33 ÷ 0.097 ≈ €3.4 — which matches the office €3.41. If CPC dropped to €0.20 with CR at 4%, CPA would increase: cheap click ≠ cheap conversion.

3. Crypixchange funnel in Telegram Ads

1. Impression in the thematic / scored channel inventory Euro-office.
2. Click (CPC-event) → landing page /exchange form on crypixchange.com with UTM.
3. Offer match: The user sees the promised pair (USDT→EUR, Crypto→Card, etc.).
4. Checkout friction: amount, rate, details, time of execution.
5. Completed exchange (CPA-event) → Action in Ads Manager + record in CRM exchanger.

For the exchanger vertical, this is closer to performance leadgen with a monetary event than to a 'channel subscription.' A detailed map of the stages is in the guide. click → conversion; crypto UA-context — in Crypto user acquisition.

4. Portfolio of creatives: what was tested

The angle matrix is built around output intensityrather than an abstract “crypto exchange”:

• Pair-led: USDT to EUR / USD, BTC to EUR / Cash.
• Rail-led: Crypto to Card, Crypto to Revolut.
• Cash-led: EUR Cash, Crypto to Cash, BTC to Cash.
• Broad: Crypto Exchange / USDT Exchange (wide message control).

Hypothesis: rail-led and clear stablecoin→fiat will give a better CPA because the post-click path is shorter and the user's expectation matches the exchanger UI. Cash-led — higher curiosity and higher friction → worse CPA. The segment confirmed the hypothesis.

5. Active winners: where is the best CPA now

After ranking by CPA, four cells remained in Active. This is the answer to the operational question 'what to focus on next'.

Active Pool Summary: 248 exchanges · spend ≈ €941.70 · blended CPA ≈ €3.80 · blended CPC ≈ €0.38 · average CTR ≈ 1.8%. USDT to EUR gives scale (102 Actions); Card — efficiency champion; Revolut — proof of high CTR without bloating CPA.

5.1 Crypto to Card — €3.41 CPA

The best slice cell. 41,885 impressions → 662 clicks → 64 exchanges. CTR 1.58% is not a record, but click→exchange CR ≈ 9.7% keeps CPA at a minimum. Conclusion: for the exchanger, 'average CTR + strong rail-intent' often beats 'high CTR + weak CR'.

5.2 Crypto to Revolut — €3.83 CPA, CTR 2.58%

The maximum CTR in the portfolio at CPA is still < €4. Small spend (€76.54) — cell in hold/scale-candidate mode: need to add n exchanges without changing the angle. Fintech-rail in creative increases the expected value of the click.

5.3 USDT to EUR — €3.87 CPA, 102 exchanges

Workhorse of the volume: 52,394 impressions, CPC €0.40, 102 exchanges. For the media plan, this is cell No. 1 by absolute contribution to Actions at CPA within the top quartile. The USDT→EUR pair is read unambiguously in the Telegram audience EU/CIS.

5.4 Crypto Exchange (core) — €4.07 CPA

Broad-angle, which is still passing the Active threshold. 62 exchanges at CPC €0.51. We hold it as control / volume filler, but not as a creative brief for new tests: it’s better to build new hypotheses from Card/Revolut/pair, not from the generic 'exchange'.

6. Stopped pool: why disabled

Below are cells with the status Stopped (rounded office values). General pattern: CPA rises along with CPC and/or click→exchange CR falls; Cash and 'to Card' in the USDT formulation without a strong match gave the worst tail.

Stoppedサマリー: ≈ 524 exchanges · spend ≈ €3.25k · blended CPA ≈ €6.20. BTC to Cash (€9.71) and USDT to Card (€9.76) tails are a classic anti-pattern: expensive clicks + few Actions. Leaving them Active would mean subsidizing an inefficient intent at the expense of Card/USDT→EUR.

7. Active vs Stopped: One Economy Table

To report to the customer, it is more correct to show 2 CPA: portfolio (with learning) and active-operating. Otherwise, 'average CPA €5.6' hides the fact that a pool of €3.8 is currently circulating. The methodology is — in ROI measurement и optimization checklist.

8. Diagnostics: CTR × CPC × CPA

@@TG Target @ @ matrix for exchanger (simplified for this slice):

• High CTR + low CPA (Revolut): scale carefully without breaking the angle.
• Medium CTR + low CPA (card, USDT→EUR): the main volume; copy the offer structure.
• Normal CTR + high CPA (part Cash / USD): problem after click or weak pair-intent — stop or redesign landing match.
• Low CTR + high CPC/CPA (USDT to Card): creative + audience misfit — hard stop.

Associated diagnostic framework — CPAを下げる方法. Key of the case: not "reduced the rate", but cut off the tail of the offers and left top-CPA Active.

9. What was done operationally (playbook)

1. The primary event was recorded: completed exchange = Action.
2. We deployed a creative matrix by rails and pairs (not a single 'Crypto Exchange' for the entire budget).
3. Collected UTM for each corner → checking Ads Manager ↔ CRM exchanger.
4. After performing min-n exchanges per cell, they were sorted by CPA.
5. Active = lower quartile CPA (here €3.41–€4.07); Stopped = the rest.
6. The budget was shifted to Active; broad Exchange was left as control, not as a growth thesis.
7. Next sprint: add n on Revolut; new tests only from Card / USDT→EUR hooks; Cash — only if the post-click flow changes.

10. GEO and case localizability

The cut was taken in @@Telegram Ads7@@-office Telegram Ads. The specific country-split is not exposed in the UI screen — therefore in the case we do not attribute CPA to one country. For GEO-adaptation:

• In EU/EEA rail offers (Card, Revolut, SEPA-like), the wording is usually stronger than Cash-P2P.
• In the CIS, USDT→local fiat and card payouts are often stronger — but a separate pilot is needed, not copy-paste €3.41.
• CPM corridor look in CPM by country; the decision on scale — only after CPA exchange and the exchanger’s unit economics (margin, limits, chargeback/fraud).

11. Compliance and moderation notes

Crypto exchanger at Telegram Ads operates with policy-safe creatives: without guarantees of an "above-market" rate, without aggressive yield language, with a real brand and a transparent website. Overall crypto allowance — can you advertise cryptoDIRECTIVES rules & restrictions. In the Crypix case, the angles are descriptive (pair / rail), without 'earn % per day' — this reduces the reject loop and allows accumulating statistics on CPA.

12. Anti-patterns, which the case clearly shows

• Consider it a win to have a low CPC at CPA €9+.
• Scale Cash-angle by CTR 1.4% without controlling Actions.
• Keep 12+ Active creatives "just in case" — blurs the budget.
• Comparing CPA exchanger with FTD CPA exchange / iGaming — different events.
• Optimize 48 hours at n<15 exchanges per cell.
• Прятать потери на обучение в одном «среднем CPA» без Active/Stopped split.

13. Checklist: repeat the method on your exchanger

• □ Primary KPI = completed exchange (not click, not session).
• □ Matrix of at least 6–10 angles: pairs + rails + 1 broad control.
• □ UTM / postback for each ad title.
• □ Min-n exchanges for kill/scale (benchmark ≥20–30 per cell, if margin allows).
• □ Ranking by CPA; Active = top quartile.
• □ Report: portfolio CPA + operating CPA.
• □ A new sprint of hypotheses only from the winners, not from the "average".
• □ Policy-safe copy before the budget is filled in scale.

14. How to read a case without self-deception

772 exchanges is enough to rank angles within this account, but too few for a global vertical benchmark. The exchanger's margin, average exchange value, and fraud rate are not disclosed — without them ROAS/ROMI from CPA cannot be derived. Use the case as solution template (what to consider KPI, how to cut the tail), and not as a promise “CPA €3.4 on any GEO”. Adjacent materials for calibrating expectations: collection of case studies, benchmarks 2026, ショートケース.

FAQ