Where should a photo product look for growth?
I connected growth drivers, regional demand, and customer value to a focused acquisition and reactivation agenda.
- My role
- End-to-end analytics
- Tools used
- SQL · Excel · Tableau
- Deliverable
- Analysis & recommendations
The problem
Activity was recovering. Leadership needed to know which users and markets were driving it, and where further investment made sense.
What I did
- Separated new-user growth, existing-user recovery, audience size, and usage frequency.
- Compared regional demand, payer economics, and feature adoption to recommend where to focus.
Existing users drove most of the recovery.
Existing users contributed 28k of the 32k increase in daily active users. The faster-growing new-user segment accounted for a smaller share of the gain.
Two views of the same 32k increase
View data & definitions
Two separate explanations of the same change; do not add the panels together. Both use focused vertical scales starting at 466k DAU. The size/frequency decomposition changes volume first at baseline frequency, then frequency at current volume; the allocation is order-dependent and is not causal.
| Lifecycle segment | November DAU | February DAU | Change | Growth rate |
|---|---|---|---|---|
| New users | 25,000 | 29,000 | +4,000 | 16.0% |
| Existing users | 455,000 | 483,000 | +28,000 | 6.2% |
Recommendation: Evaluate acquisition and reactivation together; a growth rate alone can hide where the larger contribution comes from.
Fast growth and large growth pointed to different markets.
Hainan grew 40% from a small base, adding 1.6k DAU. Beijing grew 26.2% and added 17k. The atlas lets you compare momentum, audience size, and customer value.
One geography. Three investment questions.
Beijing
- February DAU
- 82K
- Change in DAU
- +17k
- Value index
- 1.38×
The largest contribution to the net gain: 17k of 32k additional DAU.
Recommendation: Prioritize markets where scale, momentum, and value align; investigate reactivation in valuable markets where activity weakened.
Compare all reporting regions
Keep contribution and growth rate side by side
View data & definitions
| Region | November DAU | February DAU | DAU change | Growth rate | Revenue / active user index |
|---|---|---|---|---|---|
| Beijing | 65,000 | 82,000 | +17,000 | 26.2% | 1.38 |
| Guangdong | 70,000 | 75,000 | +5,000 | 7.1% | 1.16 |
| Shandong | 30,000 | 33,000 | +3,000 | 10.0% | 0.93 |
| Hunan | 22,000 | 24,000 | +2,000 | 9.1% | 0.89 |
| Hainan | 4,000 | 5,600 | +1,600 | 40.0% | 1.05 |
| Guangxi | 12,000 | 13,500 | +1,500 | 12.5% | 0.87 |
| Yunnan | 10,000 | 11,400 | +1,400 | 14.0% | 1.04 |
| Jiangsu | 35,000 | 34,000 | -1,000 | -2.9% | 1.22 |
| Shanghai | 32,000 | 30,500 | -1,500 | -4.7% | 1.40 |
| Sichuan | 38,000 | 36,800 | -1,200 | -3.2% | 1.18 |
| Northwest group | 20,000 | 19,200 | -800 | -4.0% | 0.76 |
| Other regions | 142,000 | 147,000 | +5,000 | 3.5% | 0.60 |
Ages 18–34 contributed 84% of the net gain.
That made younger adults a useful growth signal. Their higher payment rate did not mean the highest spending per payer: demand and customer value were separate questions.
Three measures, three different decisions
View data & definitions
February, using the same 20-day active-user population. Payment rate = unique payers / active users; revenue per active user = payment rate × revenue per payer. CU means illustrative currency units. Gross period revenue assumes no refunds; it excludes costs and is not lifetime profit.
| Age band | Active users | Payers | Payment rate | Revenue / payer (CU) | Revenue / active user (CU) |
|---|---|---|---|---|---|
| Under 18 | 62,000 | 744 | 1.2% | 10 | 0.120 |
| 18–24 | 490,000 | 22,540 | 4.6% | 12 | 0.552 |
| 25–34 | 770,000 | 31,570 | 4.1% | 15 | 0.615 |
| 35–44 | 465,000 | 15,810 | 3.4% | 20 | 0.680 |
| 45–54 | 266,000 | 6,916 | 2.6% | 25 | 0.650 |
| 55+ | 224,000 | 4,704 | 2.1% | 30 | 0.630 |
| Unknown | 293,000 | 7,325 | 2.5% | 17 | 0.425 |
Recommendation: Compare payment rate and value per payer before deciding how much to spend on an audience.
See age contributions and profile coverage
Keep the unclassified users in the growth story
View data & definitions
| Age band | November DAU | February DAU | Change | Share of net gain |
|---|---|---|---|---|
| Under 18 | 9,000 | 9,300 | +300 | 0.9% |
| 18–24 | 86,000 | 101,000 | +15,000 | 46.9% |
| 25–34 | 148,000 | 160,000 | +12,000 | 37.5% |
| 35–44 | 89,000 | 93,000 | +4,000 | 12.5% |
| 45–54 | 53,000 | 52,400 | -600 | -1.9% |
| 55+ | 40,000 | 38,300 | -1,700 | -5.3% |
| Unknown | 55,000 | 58,000 | +3,000 | 9.4% |
A profile can be useful without representing everyone
View data & definitions
| Profile dimension | Usable label | Unknown / unusable | Coverage |
|---|---|---|---|
| Age | 2,277,000 | 293,000 | 88.6% |
| Gender | 2,400,000 | 170,000 | 93.4% |
| Life stage | 1,800,000 | 770,000 | 70.0% |
| Marital status | 600,000 | 1,970,000 | 23.3% |
Acquisition should start with a photo-management need.
Photo viewing, albums, and intelligent retrieval suggested practical reasons to use the product. A redesign coincided with these changes, so adoption alone cannot establish a feature’s impact.
Distinguish percentage-point change from relative growth
View data & definitions
Distinct feature users / all active users in each equal 20-day window. Features overlap. A move from 1.0% to 1.5% is +0.5 percentage points, not +50 percentage points. The historical redesign makes exposure, event definitions and availability important confounders.
| Feature | November adoption | February adoption | Change (pp) |
|---|---|---|---|
| Photo preview | 43.0% | 46.5% | +3.5 |
| Albums | 15.5% | 18.0% | +2.5 |
| Smart organization | 7.3% | 9.2% | +1.9 |
| Photo search | 2.3% | 3.1% | +0.8 |
| Smart photo tags | 1.0% | 1.5% | +0.5 |
| Stories | 8.1% | 8.5% | +0.4 |
| Cleanup tools | 14.0% | 10.3% | -3.7 |
| Manual creations | 1.4% | 1.2% | -0.2 |
| Home-screen widget | 1.2% | 1.1% | -0.1 |
Connect the audience, the moment, and the product need.
- Balance acquisition and reactivation. Give the existing-user base a place in the growth plan.
- Choose markets by contribution and value. Use growth rates alongside audience scale and customer economics.
- Test need-led entry points. Match the message and timing to travel, organizing photos, or finding past memories; evaluate each opportunity.
What I delivered: An executive growth diagnosis and recommendations. Implementation uplift was not measured in this report.
Methods, definitions & source notes
The comparison uses the first 20 days of November 2022 and February 2023. Daily active users equal people active during each period × average active days ÷ 20. The new/existing groups are defined within each period.
- Age and region are overlapping views of the same DAU change; their contributions must not be added.
- Payment rate and revenue per payer use the same February population. Gross revenue is not profit or lifetime value.
- Unknown profiles remain unallocated. Grouped reporting regions are not redistributed into province-level values.
- Reopening, seasonality, marketing, and redesign may coincide; the report does not isolate their separate effects.
Original work: SQL, Excel, and Tableau. Portfolio charts were rebuilt in Python; the atlas uses projected Natural Earth geography. Original employer files, SQL, screenshots, and customer data remain private.
Download chart data