LTV и CAC на собеседовании маркетинг-аналитика

Q: LTV в РФ — какой scope?

Часто 12-24 месяца (long-term LTV). Discount rate выше из-за volatility.

Q: Free users в LTV?

Если free converts в paid eventually — counted в acquisition funnel. LTV — для paying customers.

Q: CAC drop after viral?

Yes. Referral / organic снижают blended CAC. Track отдельно.

Q: LTV / CAC vs PG (payback)?

Both. LTV/CAC — long-term profitability. Payback — cash needs.

Q: Сколько в кейсе на LTV?

Часто 5-15 минут. Calculate, interpret, recommendation.

Зачем LTV и CAC на собесе маркетинг-аналитика

LTV / CAC — фундаментальные метрики unit economics. На собесе маркетинг-аналитика кейс типичный: «оцени sustainability нашей growth strategy». Без понимания LTV / CAC ответ — слабая позиция.

Слабый ответ — «LTV = avg revenue × time». Сильный — про cohort-based LTV, payback period, ratio thresholds, segmentation.

CAC: Customer Acquisition Cost

Формула: total marketing + sales spend / new customers acquired.

Виды:

Blended CAC — для всех customers вместе
Paid CAC — только paid acquisition (excludes organic)
Channel CAC — по каналам

Calculation pitfalls:

Включать sales team salary? Если есть sales — да.
Brand spend allocation? Allocate proportionally.
Organic users исключаются из paid CAC.

Пример:

Spend на Yandex Direct $10k / month
New customers from Yandex Direct: 200
CAC = $50

LTV: Lifetime Value

Простая формула: ARPU × Gross Margin × (1 / churn rate)

Более точно (cohort-based):

Track revenue per user в cohort
Sum across cohort lifetime
Aggregate как avg per user

Формула для subscription:

LTV = ARPU × Gross Margin / Churn rate

Discount future revenue:

LTV = sum(monthly_revenue_t × (1 - churn)^t × (1 + discount_rate)^(-t))

Дисконтирование (NPV): future revenue стоит меньше today. Discount rate = cost of capital (10-20%).

LTV / CAC ratio

Standard thresholds:

< 1: убыточно
1-3: marginal, sub-optimal
3+: healthy unit economics
5+: room для accelerated investment

Why 3+?

Cover other costs (R&D, ops, overhead)
Profit margin

Payback period

Формула: CAC / monthly contribution margin.

Примеры:

CAC $100, ARPU $50, Margin 50% → monthly contribution $25 → payback 4 months
CAC $500, ARPU $20, Margin 70% → $14 → payback 36 months

Industry benchmarks:

Consumer subscription: < 6 months
SMB SaaS: < 12 months
Enterprise: < 18-24 months

Faster payback = меньше cash needed = scalable growth.

Segmentation

By channel:

Yandex Direct CAC $50, LTV $200, ratio 4
Display ads CAC $30, LTV $90, ratio 3
Organic CAC ~0, LTV $300

By geography:

Москва: high CAC, high LTV
Регионы: low CAC, lower LTV

By cohort:

2024 cohort: ratio 4
2025 cohort: ratio 3 (deteriorating?)

By segment:

B2C casual: low LTV
B2C power users: high LTV
B2B: highest LTV, highest CAC

Cohort-based LTV расчёт

WITH cohort_revenue AS (
  SELECT 
    DATE_TRUNC('month', signup_date) AS cohort_month,
    user_id,
    SUM(revenue) AS user_revenue
  FROM users u JOIN orders o USING (user_id)
  GROUP BY 1, 2
)
SELECT 
  cohort_month,
  AVG(user_revenue) AS avg_ltv,
  PERCENTILE_CONT(0.5) WITHIN GROUP (ORDER BY user_revenue) AS median_ltv,
  COUNT(*) AS cohort_size
FROM cohort_revenue
GROUP BY cohort_month
ORDER BY cohort_month;

LTV прогнозирование

Для recent cohorts (still active) — прогноз.

Methods:

Survival analysis (Cox proportional hazards)
BG/NBD + Gamma-Gamma (Pareto/NBD) — стандарт для transactional
Simple: extrapolate retention curve

Tools: lifetimes (Python lib), custom ML.

Типичные вопросы

«LTV $200, CAC $80. Хорошо?»

Ratio 2.5 — marginal. Нужно либо снижать CAC, либо повышать LTV. На long term unsustainable.

«Как ускорить payback period?»

Снизить CAC (more efficient channels).
Повысить ARPU (upsell, premium tiers).
Снизить churn (retention initiatives).
Annual prepay скидки (upfront cash).

«LTV растёт, но conversion падает. Что делать?»

Investigate: power users LTV растёт, casual users churn → average LTV up but acquisition slowdown.

«Channels: какой scale?»

Channels с ratio > 3 и low CAC volatility → scale. С ratio < 2 → pause or optimize. С high ratio но low volume → test scale up.

Частые ошибки

LTV без discount. Overestimates true value.
Blended LTV для decisions. Hides channel-level economics.
Игнор payback period. Long payback = cash burn.
CAC без sales costs. Underestimates true CAC.
Одна модель LTV. Recent cohorts отличаются от старых — нужно forecasting.

FAQ

LTV в РФ — какой scope?