Как продвигать data-driven культуру
Карьерник — квиз-тренажёр в Telegram с 1500+ вопросами для собесов аналитика. SQL, Python, A/B, метрики. Бесплатно.
Зачем это знать
Компания говорит «мы data-driven», но PM решает по intuition, CEO отклоняет A/B в пользу gut feel. Аналитик frustrated.
Senior-analyst не just delivers analyses — он changes company culture. Это отличает lead-level от middle.
На senior собесах часто: «как бы you drive data culture?».
Reality check
Большинство компаний «data-informed», не «data-driven».
«Driven» = data решает. Intuitively easy decisions — data overrides.
«Informed» = data considered. But not authoritative.
«Data-driven» часто marketing term. В practice — data одно из inputs.
Barriers
1. Stakeholder skepticism
«Эти numbers не match my intuition» → dismiss.
2. Bad prior data
«Мы смотрели metric раньше, оказался wrong» → trust gone.
3. Politics
«My feature — ship it». Data threatens narrative.
4. Lazy
A/B requires effort. Shipping «because PM said so» easier.
5. No infrastructure
No tracking, no dashboards, no DWH → analyst ограничен.
Practical approaches
1. Go к decision makers
Build relationships с PMs, product leaders, exec.
Present to them directly. Не through email — via meetings.
2. Tailor к audience
- PM: product metric, user stories, competitive benchmarks
- Exec: business impact, revenue, risk
- Engineering: technical metrics, quality
Same analysis, different framings.
3. Start small
Not «overhaul everything data-driven». Pick 1-2 decisions, data-support.
Success → credibility → expansion.
4. Speed matters
Analyze quickly. Slow → people moved on.
Fast 80% analysis > slow 100%.
5. Teach SQL
Empower others. Analyst scaling через education.
Internal SQL workshops, documentation.
6. A/B everything (when possible)
Establish culture: «новая feature? A/B».
Start on low-risk, work up.
7. Show losses от ignoring data
«Вы решили ship без теста → -5% retention ended up. Could have avoided».
Post-mortem опираясь на data.
8. Celebrate wins
Public recognition когда data drove win:
«Мы использовали A/B results → +$500k revenue». Visibility builds culture.
Common mistakes
Data snob
«Your opinion doesn't matter, только data». Alienates.
Data + context + judgment.
Over-rigorous
Every decision требует 20-page analysis → slow company.
Match rigor к stakes.
Academic pitches
«P-value 0.03, CI [1%, 7%], power 0.8» к marketing director. Confuses.
Translate: «expected revenue +$X, might be higher or lower, 95% confident это between Y and Z».
Data without recommendation
«CR 5%. Here's data». «So what?».
Always include: «рекомендация — do X because Y».
Product rituals
Weekly review
Team meets, looks at metrics. Every week.
«What moved? Why?» — habit.
Decision templates
«Before shipping:
- What's hypothesis?
- How measured?
- A/B planned?
- Success criteria?».
Force data thinking.
Post-mortems
После launch / change — analyze impact. Learn.
Experimentation platform
Infrastructure encourages experiments. Easy to launch → more launched.
Working with execs
Earn trust
Early — deliver what asked для, on time, accurate.
Over time — proactively bring insights.
Contextualize
Exec doesn't know details. «Revenue up 10%» — big или small? Compare с target, last period, industry.
Executive summary
Report: first page — summary + recommendation. Appendix — details.
Simple language
«Conversion rate» > «trial to paid probability per individual».
Scaling culture
Hire right
Each hire в команду — data-aware. Senior PMs, growth people.
Docs
Central doc: «How we make product decisions». Codify culture.
Tools accessible
DWH, BI, SQL — accessible к non-analyst.
Onboarding
New employees — «data-driven» explained.
Measure culture
Indicators
- % decisions referencing data
- Experiments running
- Data platform usage
- Self-serve analyses by non-analysts
Surveys
«Do you trust our data?». «Can you access data yourself?». Track over time.
Failure modes
«Data-driven» buzzword
Claim без reality. No rituals, no experiments, gut feel prevails.
Analysis paralysis
Everything researched → nothing shipped.
Metric gaming
Team optimizes toward numbers, forgets users.
Not using data
Data exists, dashboards built — but no one looks.
На собесе
«Company culture data-driven?»
Sincere answer:
- What rituals?
- How decisions made?
- Infrastructure?
«Work to make improvements» — если не fully data-driven.
«How would you drive culture?»
Plan:
- Start с 1-2 high-impact decisions
- Tailor communication
- Build relationships
- Teach SQL
- Establish rituals (weekly review, A/B defaults)
- Celebrate wins, learn from losses
Связанные темы
- Как презентовать результаты анализа
- Data storytelling
- Как работать с амбигуous задачами
- Как задавать вопросы стейкхолдерам
FAQ
Как быстро change culture?
6-18 месяцев обычно. Depends на leadership support.
Leadership не supports?
Harder, но возможно. Start bottom-up, build proof points.
Data-driven always better?
Нет. Balance с speed, judgment, ethics.
Тренируйте продуктовое мышление — откройте тренажёр с 1500+ вопросами для собесов.