1. Function Overview

You can use the funnel report to analyze the data of each core link in the App's ad monetization chain,

including: Launch>Get Strategy>Request Ad>Secnario>Impression Ad>Click Ad

At the same time, through the conversion rate between each core link in the ad monetization chain, you can analyze and further optimize the App's ad monetization strategy.

2. Operation Guide

To find and solve data problems through funnels, you can follow the steps below:

(1) Select the app and placement on the funnel report page

It is recommended to view data by placement. events / devices aggregated for all placements cannot locate specific problems)

(2) Switch to the number of devices or DAU to view events / devices

By calculating events / devices by DAU, you can clearly see the user churn rate of each process

(3) View the conversion rate of each process in the current placement

By checking the description of each metrics and the current metrics value, determine whether the current metrics value is abnormal, and modify the background configuration/SDK according to the reference improvement plan

(4) After the optimization operation, view the metrics change trend through day compare or trend chart to observe whether the optimization operation is effective

3. Related data description

3.1 Funnel Metrics

The funnel report includes two parts of data: Events, Conversion Rate

Events: data metrics of core events in the ad monetization funnel model, such as: Launch, Request, Imp., click, etc.
Conversion Rate: The conversion rate between core events, calculated based on the number of events or the number of devices, such as Request fill rate, Impression rate, CTR, etc.

Metrics name	Type	Statistical caliber	Explanation and suggestions
Launch	Event	Launch and successful initialization of Taku SDK	Due to differences in statistical caliber and data reporting logic, Taku launch data differs from mediation network statistical SDK data
Sessions/DAU	Conversion rate	Sessions/startup devices	1. The average number of times a user launches an app in a day. 2. The impact of different app versions on user usage frequency can be analyzed through Sessions/DAU. When the data is abnormal, it is recommended that you analyze it in combination with the crash information (you can use mediation network services to collect app crash information)
Get strategy	Event	App requests placement strategy from Taku server	When the number of devices for obtaining strategy is less than the number of devices for launch, it is recommended that you add ad estimated loading logic in appropriate scenarios (you can also consider using Taku's fully automatic loading mode to simplify code access logic, View access instructions)
Request	Event	Number of times the app sends ad requests to Taku	-
Request Fill	Event	After the app sends an ad request to Taku, Taku returns that the ad request is successfully filled	When the number of tequest fills is significantly less than the request, you can consider adding an ad source with no reserve price as a backup in the WaterFall configuration of the aggregation management
Request Fill Rate	Conversion rate	Request fill / Request	For the problem of no filling of placements or ad sources, you can query the reasons for no filling through log analysis, View log analysis instructions
Scenario	Event	Number of times users reach scenario	The app needs to call Taku's entryAdScenario interface to perform statistics. Refer to Taku SDK documentation to add the code for reaching the scenario
Scenario Reach Rate	Conversion rate	The number of users reach scenario/App launch	1. When the user scenario, the app needs to call the entryAdScenario method of Taku to count. Refer to the Taku SDK documentation to add the code for the scenario 2. The arrival rate of different scenarios can reflect the rationality of the scenario design. By comparing the scenario reach rate horizontally, the scenario design of the app can be further optimized
Ad Ready Rate	Conversion Rate	The proportion of ads that are ready when users arrive at the scenario	When the ad Ready rate is low, it is recommended that you add ad estimated loading logic in appropriate scenarios (you can also consider using Taku's fully automatic loading mode to simplify the code access logic, view access instructions)
Call isReady	Event	App calls Taku's isReady interface	-
Call isReady Rate	conversion rate	The percentage of Taku isReady interface returning True	The problem of low isReady success rate can be located in the following ways: 1. Determine whether the scenario of calling isReady is reasonable and whether it is too frequent 2. Add ad estimated loading logic in appropriate scenarios (you can also consider using Taku's fully automatic loading mode to simplify the code access logic, View access instructions)
Call Show	Event	App calls Taku's Show interface (native and banner ads are to trigger Taku impression logic)	If you use an scenario, you need to pass in the scenario ID when calling the Show method,Refer to Taku SDK documentation to add scenario ID
Call Show Rate	Conversion rate	The number of Taku isReady interface returns True / the number of Taku isReady interface is called	1. The percentage of users who are triggered to watch ads after reaching the scenario. 2. For rewarded videos, if the call show rate is low, you can consider increasing incentives to stimulate users to watch ads. The ad triggers of other ad types are handled by the app. You can consider optimizing the logic of triggering ad impression.
Impression	Event	From 0:00 on March 7, 2023 (UTC+8), the statistical caliber will be adjusted from "the number of times Taku calls the network SDK to start impression" to "the number of callbacks Taku receives from the network SDK for successful impression"	-
SS Rate	Conversion rate	Impression / Call Show	For impression failure issues of placements or ad sources, you can query the reasons for impression failure through log analysis, View log analysis instructions
Impression API	Conversion rate	The number of ad impressions that Taku pulls from the network through the API. Due to differences in statistical caliber, network packet loss and other factors, there may be some differences between Taku impression and network impression.	Impression API data is based on the network Report API or upload mediation network data. It is recommended that you follow the document Complete the Report API connection of each network or Manually upload mediation network data
Impression Gap	Conversion rate	(Impression- Impression API）/ Impression API	Differences between Taku statistics and network statistics. Please refer to 2. Common Problems in Funnel Analysis for troubleshooting
Click	Event	Taku Ad Click	-
CTR	Conversion rate	Click / Impression	Taku statistics of ad click rate

Note:
1. App launch data is overall data of the app dimension, without distinguishing between placements and scenarios
2. Impression API does not support segmentation by scenario, channel, SDK version, app version, and does not support viewing of Devices and Events / Devices
3. Impression gap does not support segmentation by scenario, channel, SDK version, app version, and does not support viewing of Devices and Events / Devices

3.2 Funnel Report

(1) Funnel Report provides three different data dimensions:

Data Dimension	Statistical Caliber
Times	Calculate by the number of times an event occurs
Devices	Calculated according to the number of devices (number of users) where the event occurred
Events / Devices	Statistical caliber (number of times / number of devices) 1.Data that scenario and impression API needs to complete code integration or configuration as required to be impressioned normally

(2) Data that scenario and impression API needs to complete code integration or configuration as required to be impressioned normally

5. FAQ

(1) App Launch data is low

Due to differences in statistical caliber and data reporting logic, Taku Launch data differs from mediation network statistical SDK data.

(2) Get Strategy more than App launch

① The events of Get Strategy more than App launch: This is a normal phenomenon. Since an app will have multiple placements when it is launched, and different placements will independently request placement strategies from Taku, the number of strategies acquired will be more than the number of Launches.

② The devices of Get Strategy more than App launch: An app may have multiple days in its life cycle, but the data gap generated by such a situation is relatively small. When the DAU is small and the number of users across days is large, the data gap may be more obvious.

(3) Request more than Get Configuration

① Request is greater than Get Configuration: This is a normal phenomenon. Since Taku SDK caches the strategy, multiple requests can be initiated during the cache validity period, so the number of requests will be greater than the number of configuration acquisitions.

② Request Devices is greater than Get Configuration Devices:

For splash ads, this situation will occur if the first splash screen configuration is used.
When this situation occurs for other ad types, it is possible that some users are using the app while disconnected from the Internet.

(4) isReady Calls The Device is greater than Request Device

When the number of devices calling isReady is greater than the number of devices requesting traffic, it means that some devices have not triggered ad requests. At this time, you can optimize from the following two aspects:

① Determine whether the scenario for calling isReady is reasonable and whether it is too frequent. You can adjust the logic for calling isReady appropriately

② Determine whether the ad request logic is reasonable. You can consider adding ad estimated loading logic in appropriate scenarios (you can also consider using Taku's fully automatic loading mode to simplify the code access logic. See the access instructions)

(5) The Scenario Reach Rate is low

The scenario reach rate is low, which is generally due to unreasonable scenario design or too deep ad entry settings, so users cannot reach the corresponding scenario. At the same time, by comparing the arrival rates of different scenarios horizontally, you can further optimize the scenario entry design of your app.

(6) Low Request Fill Rate

The following aspects can be used to optimize the low request fill rate:

① When the traffic fill number is significantly less than the request, you can consider adding a no-reserve ad source as a backup in the WaterFall configuration of aggregation management

② For splash ads, you can consider adding a backup ad source. See the splash screen backup instructions

③ At the same time, you can use log analysis to query the reasons why the placement or ad source is not filled. See the log analysis instructions

(7) Low Ad Ready Rate

When the ad ready rate is low, you need to analyze it together with the request fill rate:

① High request fill rate but low ad ready rate: This means that there is no ad to impression when the user scenario. It is recommended that you add ad estimated loading logic in appropriate scenes (you can also consider using Taku's fully automatic loading mode to simplify the code access logic, see the access instructions).

② Low request fill rate and low ad ready rate: This means that there is a problem with the WaterFall configuration. You can consider adding a no-bottom-price ad source to the WaterFall configuration as a backup. At the same time, you can use log analysis to query the reasons why the placement or ad source is not filled. See the log analysis instructions.

③ For splash ads, you can consider adding a backup ad source. See the splash backup instructions.

(8) Low Call Show Rate

① For rewarded videos, if the call show rate is low, you can consider adding rewarding content to stimulate users to watch ads.

② For other ad types, ad triggering is handled by the app. You can consider optimizing the logic for triggering ad impression.

(9) Call Show Rate exceeds 100%

① Call show rate = call show / scenario. You need to make sure that the app has correctly added scenario and the code for Call Show.

② Scenario: When the user arrive scenario, the app needs to call the entryAdScenario method of Taku to count the scenario. Refer to the Taku SDK documentation for instructions on adding the code for reaching the scenario

③ Calling Show: When the app display an ad, it needs to pass in the scenario ID when calling the Show method. Refer to the Taku SDK documentation for instructions on adding the scenario ID

(10) Show rate is low

You can view the detailed reasons for the impression error through log analysis. Please select the abnormal placement. The ad impression may fail due to cache expiration or SDK access problems.

(11) Low SS rate

① For rewarded video and interstitial ads, you can use log analysis to query the reasons for the failure of placement or ad source impression. See the log analysis instructions.

② For native, banner, and splash ads, you need to check whether there is a problem with the impression ad code logic.

(12) Impression Gap exceeds 15%

Impression Gap problems are usually caused by a variety of reasons. You can try to start from the following aspects:

① Impression API data is based on the network Report API connection or uploading mediation network data. It is recommended that you complete the Report API connection of each network or manually upload mediation network data according to the document.

② There will be a delay when the Report API pulls the impression API data. For the latest data of each day, you can view the data at the following time nodes:

utc+8: delayed by 1 day, updated at 18:00 pm (utc+8)
utc+0: delayed by 1 day, updated at 18:00 pm (utc+8)
utc-8: delayed by 2 days, updated at 1:30 am (utc+8)

③ If Report API docking has been completed or mediation network data has been uploaded and there is still a impression Gap, you can use the following methods to troubleshoot

Too many impression APIs:

Check whether the mediation network placement IDs are reused (the placement IDs of the same mediation network network are configured on different ad sources of Taku)?

Impression API is too few:

All ad sources of a certain ad type on a certain network have impression gap problems: It may be a code integration problem. You can refer to 10. Methods for low SS rate to troubleshoot the reasons for impression failure.
Some ad sources of a certain ad type on a certain network have impression gap problems: The impression of this ad source is determined to be invalid by the network. You can try to communicate with the network.