Assignment strategies
liftstack offers three traffic allocation strategies for distributing recipients across variants.
Equal split
Each variant receives an equal share of traffic. This is the fairest approach and recommended for your first campaigns. Available on all tiers.
Manual allocation
Set custom percentages for each variant. Useful when you want to limit exposure to a risky variant or when regulatory requirements dictate allocation. Available on all tiers.
Thompson Sampling (Growth and Scale tiers)
When you have tested the same snippet variants across multiple campaigns, liftstack can use historical performance data to send more traffic to the variants that have been performing well, while still sending some traffic to underperforming variants to make sure nothing is being missed. This is called Thompson Sampling.
For example, instead of a 33/33/33 equal split across three variants, liftstack might recommend a 60/25/15 split based on past performance. The likely winner gets more traffic (fewer wasted exposures), while alternatives still get enough to confirm whether they have improved or the leader has slipped.
Thompson Sampling does not bias the test. The system still tracks performance for every variant and runs the full statistical analysis. The unequal allocation actually makes the test more efficient, because you reach conclusions faster with more recipients exposed to the likely best variant.
Transparency panel
When liftstack recommends an allocation, you will see a transparency panel showing:
- The recommended traffic split
- Why the system is recommending this split (citing historical conversion rates)
- Three options: Accept, Adjust Manually (drag sliders), or Use Equal Split
Smart Allocation Uplift
When a campaign uses Thompson Sampling, the report shows an extra metric: the additional conversions captured by the smart allocation compared to what an equal split would have produced. This isolates the value of the allocation strategy from the value of testing itself.
New variant handling
New variants (those that have never appeared in a completed campaign) receive a guaranteed minimum of 20% of traffic on their first campaign, regardless of what Thompson Sampling would recommend. This prevents established variants from starving newcomers of exposure.
Historical data decay
liftstack applies a recency decay to historical data: performance from campaigns 60 days ago counts half as much as recent campaigns, and very old data fades away almost entirely. This ensures the allocation reflects current audience preferences, not stale data.
Built-in guardrails
To prevent any single variant from being starved of traffic or over-exposed:
- The control always receives at least 10% of traffic
- No variant drops below 5%