What acquisition plan can meet the growth target?
I translated an active-user target into acquisition schedules and compared the timing, retention, and cost assumptions behind them.
- My role
- Modeling & decision support
- Tools used
- SQL · Excel · Python
- Deliverable
- Modeling & decision support
The problem
A DAU target did not tell the business how many users to acquire, when to acquire them, or which channels could justify their cost.
What I did
- Connected acquisition cohorts to future activity through retention.
- Compared schedules, interim milestones, channel economics, and downside assumptions.
A plan can meet the final target and still fall short earlier.
Both default schedules reach 560k DAU from a 420k baseline. Ramping uses fewer acquisitions, but delivers fewer active user-days and misses earlier milestones. Try changing the assumptions.
Same final target.
Different paths to get there.
At equal unit prices, ramping uses fewer additional acquisitions to meet the final target. It also delivers fewer active user-days and misses earlier illustrative milestones.
| Measure | Even | Ramp |
|---|---|---|
| Final DAU | 560K | 560K |
| Additional acquisitions | 3.11M | 2.84M |
| Acquisition spend (CU) | 4.98M | 4.54M |
| Additional active user-days | 32.83M | 27.8M |
| Quarter 1 DAU / floor | 486.32K / 469K · Met | 464.81K / 469K · Missed |
| Quarter 2 DAU / floor | 515.99K / 504K · Met | 493.82K / 504K · Missed |
| Quarter 3 DAU / floor | 539.61K / 532K · Met | 525.74K / 532K · Missed |
| Quarter 4 DAU / floor | 560K / 560K · Met | 560K / 560K · Met |
Recommendation: Choose the interim milestones before optimizing the final date. Fewer acquisitions do not establish a realized budget saving.
How retention connects acquisition to future activity
Each day’s acquired users contribute to future activity according to cohort age. Add those contributions to the baseline to build each schedule’s DAU path.
An assumed retention curve turns acquisition into future activity
View data & definitions
| Day | Original cohort | Active on day | Simulated activity rate | Assumed rate |
|---|---|---|---|---|
| 1 | 20000 | 6172 | 30.9% | 31.0% |
| 7 | 20000 | 2724 | 13.6% | 13.4% |
| 14 | 20000 | 2017 | 10.1% | 10.0% |
| 30 | 20000 | 1431 | 7.2% | 7.2% |
| 60 | 20000 | 1115 | 5.6% | 5.3% |
| 90 | 20000 | 918 | 4.6% | 4.5% |
The model holds the existing baseline flat and uses a shared retention curve. Test whether those assumptions transfer to new cohorts and channels.
Higher customer value can still be the worse acquisition choice.
Partner referrals generate less revenue per user than paid discovery, but have a higher revenue-to-cost ratio. The denominator changes the channel ranking.
Compare expected revenue with the costs needed to earn it
View data & definitions
Each point represents a different channel. CU denotes currency units. Cost includes acquisition and modeled bandwidth/storage service over the same 12-month horizon; other operating costs, refunds and uncertainty are not modeled. Points above the diagonal cover these modeled costs, not necessarily all business costs.
| Channel | 12-month revenue / user (CU) | Acquisition / user (CU) | Service / user (CU) | Revenue / modeled cost |
|---|---|---|---|---|
| Paid discovery | 3.20 | 3.40 | 0.80 | 0.76× |
| Partner referrals | 1.40 | 0.35 | 0.45 | 1.75× |
| Content sharing | 2.10 | 0.95 | 0.75 | 1.24× |
| Adult student communities | 5.90 | 3.80 | 1.05 | 1.22× |
Recommendation: Compare channels over the same revenue horizon and cost scope, alongside capacity and customer quality.
A price premium can erase the ramp plan’s spend advantage.
Weaker retention raises the acquisition requirement. A price premium for the ramp schedule can erase its spend advantage. The matrix makes those dependencies visible.
Retention and price assumptions change the modeled acquisition budget
View data & definitions
The final target stays 560k DAU from a flat 420k baseline. All cells are deterministic scenarios, not probabilities or confidence intervals. Spend is for incremental acquisition only; no cost saving has been observed.
| Retention scenario | Ramp unit-cost premium | Modeled ramp spend (CU) |
|---|---|---|
| weaker | 0.0% | 7,129,196 |
| weaker | 5.0% | 7,485,656 |
| weaker | 10.0% | 7,842,116 |
| weaker | 15.0% | 8,198,575 |
| weaker | 20.0% | 8,555,035 |
| base | 0.0% | 4,543,310 |
| base | 5.0% | 4,770,475 |
| base | 10.0% | 4,997,641 |
| base | 15.0% | 5,224,806 |
| base | 20.0% | 5,451,971 |
| stronger | 0.0% | 2,925,037 |
| stronger | 5.0% | 3,071,288 |
| stronger | 10.0% | 3,217,540 |
| stronger | 15.0% | 3,363,792 |
| stronger | 20.0% | 3,510,044 |
Choose a feasible path and make its assumptions inspectable.
- Agree on success. Set interim activity milestones and the cost scope before choosing an acquisition schedule.
- Compare consistent economics. Include acquisition and service costs over the same revenue horizon.
- Stress retention and price. Revisit the plan when the cohort mix or acquisition conditions change.
What I delivered: Python planning models and Excel-based channel evaluation to support resource discussions. The historical LTV dashboard was still under construction.
Model assumptions, validation & source notes
The model adds baseline activity to each earlier acquisition cohort’s expected activity. It assumes a flat baseline, available acquisition capacity, and a retention curve shared across schedules.
- Four quarters each contain 90 days. The retention curve is an assumption, not an out-of-sample forecast-accuracy claim.
- The cost slider applies a hypothetical premium to the ramp schedule only; it does not estimate how advertising prices respond to spend.
- Channel economics use 12-month gross revenue divided by acquisition plus modeled service cost. That ratio is not net ROI or full business profit.
- Scenarios are deterministic, not probabilities or confidence intervals. Modeled spend differences are not observed savings.
The browser calculations are checked against independent Python references. Original employer files, SQL, and datasets remain private.
Download the default scenarios