Frequently asked questions
Which Lemon AI capabilities are available?
Lemon AI has four public capabilities: Creative Analytics, Attribute Analysis, Creative Generation, and Cohort Prediction. The first three are included in the self-serve plan. Cohort Prediction is configured and enabled with the Lemon team.
Which ad networks are supported for Creative Analytics?
Lemon connects AppLovin, Meta, Google Ads, Unity Ads, TikTok, Mintegral, Kayzen, and Apple Search Ads. Each network provides different metrics and attribution windows, and Lemon shows what is available for every connected account.
How does Lemon estimate AppLovin asset outcomes?
AppLovin gives Lemon the impressions, clicks, and cost for each asset. Lemon combines them with matching outcome rates to estimate installs, purchases, revenue, and D0 to D7 performance. One AppLovin ad can contain several assets, so use those estimates to compare assets rather than adding them together.
Does N/A mean the value is zero?
No. N/A means the network does not provide that value. Zero means the metric is available and its value is zero. Lemon also marks cohort values that are still in progress.
Does Lemon replace an MMP?
No. Creative Analytics reads reporting data from connected ad networks and works alongside an MMP such as AppsFlyer or Adjust. Cohort Prediction can use pseudonymous tracker or MMP data as part of its configured input.
What does Attribute Analysis support?
You can define AI-generated or manual attributes for images, videos, and ad copy. Lemon can return a category, number, or yes-or-no value, then compare it with the performance metric you choose.
What can Creative Generation produce?
Creative Lab generates image and video ad assets from a prompt, product URL, dimensions, and reference creatives. Approved work can be downloaded, resized, used as the start of a new chat, or pushed to supported connected ad accounts. Video delivery requires a YouTube connection.
How is Cohort Prediction validated?
Before your dashboard goes live, Lemon predicts the value of past cohorts without looking at their eventual results. It then compares the forecast with what those cohorts earned. The D365 backtest range is 5 to 10% mean absolute percentage error.
What data is required for Cohort Prediction?
Setup uses pseudonymous tracker or MMP identifiers, user actions, revenue events, acquisition dimensions, and relevant static device or attribution features. Names, email addresses, and phone numbers are not required. Exact history and volume requirements are determined during setup.
What is your pricing model?
Creative Analytics, Attribute Analysis, and Creative Generation cost $279 per month per app with unlimited users and connected ad networks. Cohort Prediction costs $1,399 once per account for setup, followed by $679 per month per app.
How does the free trial work?
The trial lasts 7 days, requires no card, and covers up to 2 apps. It includes 3,000 generation credits and 1,200 attribute credits. Trial credits expire when the trial ends.
Does Lemon change bids, budgets, targeting, or publish creatives automatically?
No. Your team decides what to change. Lemon sends a generated creative to a connected network only after someone chooses that action. Reporting API and MCP access are read-only.
Is Lemon’s cohort prediction technology patented?
Yes. The per-user prediction approach is covered by U.S. Patent 12,014,278 B1, issued June 18, 2024.