Author:Head of Campaign Diagnostics TG Target · pilot/recovery · iGaming, crypto, fintech EU/CIS · FAILURE MAP audits.
"Doesn't work" in Telegram Ads— diagnosis without specifying the failure layer. TG Target (EU/CIS): the campaign breaks down on one of the six layers FAILURE MAP - delivery, pre-click, post-click, measurement, strategy, policy. There is delivery, but there are no conversions - the problem is the funnel or tracking, not “Telegram is not suitable.” CTR high, FTD no - message-market disconnect. CPA “ok in the office”, CRM unprofitable - measurement fail. Without layer-diagnosis you change GEO instead of fix after click.
The main thing from the article
- “Doesn’t work” almost always = specific FAILURE MAP layer, not the entire channel.
- Clicks without FTD - after a click or measurement; not targeting “at all”.
- CTR without main KPI CR - empty optimization.
- Pilot €6–20k / 14–28 days – minimum for valid withdrawal (EU/CIS).
- Recovery: layer audit → channel cut → tracking → creative wave, not a new cabinet.
What does “campaign not working” mean?
Direct answer: the formulation without KPI and the metrics layer is meaningless. “Not working” can mean: no delivery; there are clicks, no registration; registration yes, no FTD; FTD yes, ROI <0; scale does not hold CPA - five different problems and five different fixes.
Entities: Telegram Ads, Telegram Ads Manager, performance marketing, media buying, Telegram Ads, FTD, CTR, CPM, conversion funnel, Match Rate, Channel Fit, Primary Telegram Ads, Telegram Ads, ROAS. Audit process -Campaign audit; tracking -Analytics and tracking.
Unique block: TG Target FAILURE MAP - six fault layers
Typical SEO articles give a list of “10 reasons.” FAILURE MAP specifies the diagnostic order: which layer to check first for a specific symptom.
| Layer | Symptom | Root cause (typical) | First fix |
|---|
| F1 — Delivery | No impressions / spend does not go away | Bid, budget, policy hold, GEO inventory | Bid step, budget floor, moderation |
| F2 — Pre-click | CTR <0.3%, few clicks | Channel mismatch, weak hook, ad fatigue | A set of channels, a wave of creatives |
| F3 - After click | There are clicks, FTD/deposit no | Landing CR, bot flow, offer mismatch | Funnel audit, focus on the main KPI |
| F4 — Measurement | “CPA ok” in Ads, CRM is unprofitable | Match rate <95%, wrong KPI proxy | TRACK STACK, FTD postback |
| F5 - Strategy | Pilot “fail” in 5 days for €800 | Low budget, wrong GEO tier, no gates | Pilot redesign, success criteria |
| F6 — Policy | Failure loop, unstable delivery | Vertical restrictions, aggressive copy | Moderation-safe angles |
Rule: Do not raise F5 (change GEO) until F3 and F4 are eliminated. 60% “Telegram does not work” - F3+F4 (TG Target audits EU/CIS).
How to Diagnose: Step-by-Step Process
- Fix main KPI:FTD, deposit, KYC - not registration by default.
- F1 check:impressions, spend pace, policy status - delivery alive?
- F2 check:CTR vs vertical benchmark; channel-level CTR split.
- F4 check (early!):match rate, click→session - before optimizing creatives.
- F3 check:stage CR funnels - click→reg→FTD; where is the leak.
- Disabling channels:utm_term / channel_id - who is pulling CPA up.
- Stop/fix/scale solution:according to the main KPI + economics, not CTR.
KPI→Delivery→Tracking→Funnel→Channels→Solution
Why is it important to distinguish between layers (and not “bad channel”)
- Waste budget:shift GEO at F3 burns €5–15k without funnel fix (TG Target typ.).
- False negative:Telegram “doesn’t work” on the board deck, Meta gets the budget - although the problem was in tracking.
- Scale trap:+100% spend on F4 (false CPA) scales the loss.
- Churn Agency:no FAILURE MAP status - the client and the agency speak different languages.
F1–F2: delivery and pre-click - when the campaign doesn’t work
| Symptom | Probable Cause | Check | Fix |
|---|
| 0 impressions in 48 hours | Bid below floor, narrow GEO | Bid vs CPM benchmark | +15–25% bid step |
| Spend <€50/day with cap €500 | Fine inventory, schedule | Channel Set Size | Expand Core+Explore |
| CTR 0.15–0.25% | Explore noise channels | Channel-level CTR | CFS cut, Core pack |
| CTR fell for 7 days in a row | Fatigue of creatives | Frequency proxy | Wave of creatives: 3–5 new |
Channel selection -Channel selection strategy.
F3–F4: after click and measurement are the main “invisible” killers
Direct answer: if the clicks are stable, but the main KPI is not - in 70% of cases TG Target finds F3 (funnel) or F4 (tracking), sometimes both.
| Pattern | Layer | Signal TG Target | Fix |
|---|
| CTR 1.2%, FTD 0 | F3 | Click→reg ok, reg→FTD 0 | Payment flow, offer, KYC friction |
| Reg CPA “on target”, FTD CPA ×2 | F3 | Vanity KPI optimization | Switch primary to FTD |
| Ads CPA ok, CRM CPA +40% | F4 | Match rate 72–88% | Fix UTM/KYC redirect |
| Blended CPA ok, 70% channels red | F3+F4 | No channel muting | utm_term + channel pause |
F5–F6: strategy and policy - when it “doesn’t work” before the start
- Underpowered pilot:€500, 3 days, 1 ad - statistically invalid; not a verdict on Telegram.
- Wrong GEO tier:Tier-1 CPM to Tier-3 economics - CPA always “does not converge.”
- Meta expectations:D1 ROAS gate — Telegram rarely gives high risk on day 1.
- Policy:iGaming/crypto reject → watered-down creative → CTR/CR collapse. See moderation guides.
Mini-case: recovery iGaming (EU/CIS)
Client: “Telegram doesn’t work” · 21 days · €18k spend · registration CPA “ok” · FTD ROI −35%.
| Nakhodka | Layer | Action | Δ after 14 days |
|---|
| Match rate 81% | F4 | UTM via deposit redirect | 81% → 96% |
| Opt by registration | F3 | Main KPI → FTD | FTD CPA −28% |
| 3 channels = 62% spend, 0 FTD | F2+F3 | Disabling channels | spend efficiency +40% |
| Curiosity hook CTR 2.1% | F2 | Context/Benefit wave | FTD CR +0.9 p.p. |
Result of recovery 28 days: FTD CPA −34%, ROI D30 +12%. Telegram started working without changing GEO - the fix layer was changed.
Practical solutions: stop vs fix vs pivot
Stop - when to turn it off
Main KPI 0 with 300+ clicks after funnel review; match rate <85% without fix plan 7 days; CPA >2× target 14+ days post-fix; policy ban imminent. Don't stop by CTR alone.
Fix - when to fix it on the spot
Delivery ok; tracking fixable; ≥20% channels in CPA target; after clicking CR below benchmark, but trend ↑ after funnel change. Recovery budget: 50% pilot for 14 days.
Pivot - when to change the hypothesis
Core pack (15+ channels) is exhausted; FTD CR stable <50% benchmark after 2 funnel iterations; economics GEO does not converge with the correct LTV. Pivot = GEO/offer, not “new office same setup”.
Metrics: what to look for when diagnosing
| Metrics | Vanity / Working | Gate "works" |
|---|
| CTR | Vanity without FTD CR | Context + совпадение с каналом |
| Registration CPA | Vanity for iGaming/crypto | FTD CPA primary |
| CPM | Vanity for ROI | Channel CPA vs target |
| Match rate | Action - mandatory monitoring | ≥95% |
| FTD CPA / ROAS D14 | Action - primary | Vs target + LTV |
| Channel-level CPA | Action - optimization | Cut bottom 30% |
Comparison: “doesn’t work” in Telegram vs Meta vs Google
| Channel | Typical false reason | Real common reason | Time limit until a reasonable conclusion |
|---|
| Telegram Ads | "No audience" | After click + tracking | 14–28 days pilot |
| Meta Ads | "The pixel is broken" | Fatigue of creatives, CAPI gap | 7–14 days |
| Google Ads | "Dear CPC" | Intent mismatch, LP quality | 7–21 days |
| Affiliate | «Fraud» | Offer cap, geo scrub | Deal-dependent |
Channel Comparison -Telegram vs Google vs Meta.
Typical diagnostic errors
- The output “does not work” for days 3–5 and €<2k spend.
- Optimization for CTR or registration for FTD-business.
- Ignore match rate - CPA in Ads Manager as truth.
- Blended optimization without channel cut.
- Shift GEO to funnel/tracking audit.
- Copy Meta structure 1:1.
- There are no written pilot success/fail criteria.
- Policy issue is interpreted as “channel fail”.
Checklist: is the campaign really “dead” or fixable?
- Основной KPI и целевой CPA задокументированы.
- Pilot ≥€6k media, ≥14 days, ≥200 clicks per adset.
- F1 delivery confirmed (spend, impressions).
- Match rate measured - if <95%, F4 fix attempted.
- Funnel stage CR vs benchmark - the leak has been localized.
- Channel-level CPA — bottom 30% paused.
- ≥1 creative wave after basic level.
- STOP/FIX/PIVOT decision on main KPI, not CTR.
FAQ
Why don't advertising campaigns work in Telegram Ads?
In 80% of cases TG Target (EU/CIS audit) the reason is not a “bad channel”, but a chain: incorrect channel fit, weak post-click CR, broken tracking or optimization based on empty metrics (CTR, registration Telegram Ads). Telegram Ads - contextual performance: without a funnel, attribution and channel-level cut, the campaign looks “dead”, although there is delivery and clicks. Diagnostics - by layers FAILURE MAP, not by one digit.
Telegram Ads does not give conversions - what to check first?
TG Target order: 1) delivery (are there clicks and spend); 2) tracking match rate ≥95%; 3) click→session rate; 4) main KPI CR (FTD/deposit, not registration); 5) channel-level CPA. If there are clicks, but no conversions, the problem is after the click or measurement, not targeting “at all.” See funnel audit and TRACK STACK before shift GEO or budget.
High CTR, but no sales - what's the reason?
CTR measures pre-click; CPA/ROAS - after clicking. A high CTR with a curiosity-hook or irrelevant channel gives a cheap click and an expensive FTD. TG Target typ.: CTR 1.8% + FTD CR 0.4% = CPA 3x higher target vs CTR 0.9% + FTD CR 1.6%. Fix the message-market fit of the creative and landing, not the bet. Channel-level cut will show where CTR is “empty”.
Why doesn't Telegram Ads "pay off" with normal CPM?
CPM — cost of reach; ROI - function CR, LTV and margin after click. Low CPM on Explore channels often = low intent → high CPA. Plus: ROI is considered without TCL, without LTV or by registration instead of FTD - it seems that it “doesn’t pay off”. TG Target: compare channel CPA vs target and LTV:CAC D30, not CPM vs Meta.
How much budget do you need to understand if Telegram Ads works?
Pilot TG Target (EU/CIS): €6–20k media + €2–5k TCL, 14–28 days, min 200–500 clicks on primary ad set before kill. Less is statistically noise. One ad + €500 is not a test of the channel, but a test of patience. Gate: CPA/FTD trend 7+ days + match rate ≥95%. Early output “doesn’t work” on day 3 is a common mistake founders make.
Could a campaign not work due to moderation?
Yes. Failure of the loop, softening of the creative to “neutral” text, drop in CTR and channel fit - a typical iGaming/crypto/adult scenario. The campaign “is there”, but the delivery is unstable or the creative does not convert. TG Target: separate policy pre-check before media; do not confuse policy fail with “channel is not suitable”. See moderation guide for vertical.
Telegram Ads vs Meta: why did it work there but not here?
Different model: Meta - interest/lookalike + pixel; Telegram - channel context + weaker native conversion stack. Transferring Meta creatives and landing without adaptation results in disconnect. Waiting for D1 ROAS as in Meta is a methodological error. TG Target: pilot Telegram with channel sets and TRACK STACK, not copy-paste Meta campaign structure.
When to stop a campaign and when to fix it?
STOP: match rate <85% without fix plan; main KPI 0 at 300+ clicks; policy ban risk; CPA >2× target 14+ days after funnel fix. FIX: CTR ok, after clicking CR low; channel split shows 20% of channels with CPA in target; tracking gap. TG Target: don't kill by CTR alone; kill by main KPI + economics after min-n events.
Why does the agency’s campaign “not work” from the first month?
Telegram pilot — learning phase 2–4 weeks: channel test, creative waves, tracking stabilize. An agency without a pre-launch audit, without a TRACK registry and without written gate criteria looks “ineffective” before fixing. TG Target: success/fail pilot commit to spend; the client sees FAILURE MAP status weekly. Without this, there is a conflict of expectations, not performance.
How to restore a “dead” Telegram Ads campaign?
Recovery mode TG Target: 1) audit FAILURE MAP - determine the layer; 2) pause blended optimization; 3) channel cut — leave the Core pack with the best channel CPA; 4) fix tracking to match ≥95%; 5) creative wave with Context/Benefit hooks; 6) relaunch 50% budget 14 days. Not a “new office” without diagnostics - repeat the same mistake.
TG Target FAILURE MAP, recovery manual and pilot gate benchmarks are current as of August 2026 (EU/CIS). Thresholds CPA/FTD adapt to the product; keep the order of layers F1→F6.