performance-report

/performance-report — VK Ads statistics

Algorithm

1. Call list_campaigns (optionally status: "active") to collect campaign ids.
2. Call get_statistics with object_type: "campaigns", ids = those ids (1–200), period: "day" for a trend or "summary" for one aggregate row, plus date_from / date_to in YYYY-MM-DD. period: "day" requires both dates and the range must be at most 92 days. metrics selects the counter set (all, base, events, video, …).
3. Compute the derived metrics yourself — VK returns raw counters, and impressions are called shows: CTR = clicks / shows × 100, CPC = spend / clicks, CPM = spend / shows × 1000. Amounts are already in the cabinet currency — do not divide by 100.
4. To drill into a weak campaign: list_ad_groups with campaign_ids and get_statistics with object_type: "ad_groups", then list_ads with ad_group_ids and get_statistics with object_type: "banners". That last level is where a bad creative shows up.
5. Call get_account so the report ends with runway, not just spend.

Response format

## VK Ads — 2026-08-01 … 2026-08-31

| Campaign | Shows | Clicks | CTR | Spend | CPC |
|----------|-------|--------|-----|-------|-----|
| Осенняя распродажа | 1,204,882 | 8,412 | 0.70% | 84,120 ₽ | 10.00 ₽ |
| Ретаргет | 210,004 | 3,880 | 1.85% | 21,340 ₽ | 5.50 ₽ |

**Total spend**: 105,460 ₽
**Balance left**: 84,000 ₽ — about 25 days at the current rate.

### Weak spot
"Осенняя распродажа" banner 991205 — CTR 0.11%, 12,400 ₽ spent, 9 clicks.

Notes

• object_type accepts campaigns (= ad_plans), ad_groups, banners and users. users gives the account-level roll-up.
• period is only day or summary — there is no week / month / overall; aggregate daily rows yourself when the user wants weekly buckets.
• Statistics take at most 200 ids per call. Batch larger accounts.
• get_account reads /user.json; the balance field needs the OAuth scope read_payments.