Alerts

Create customized alerts and get notified when anomalies are affecting performance of your monetization partners. 

Getting started

1. Create your account. Learn more.

2. Upgrade to Pack M, L or XL. Learn more.

Feature overview

1. Main Controller

6. Alert Viewer

5. Schedule Controller

2. Event Controller

3.Benchmark Controller

4. Trigger Controller

Feature Walk-through

1. Main Controller. Here's where you can name your alert and select Filters for the alert. Filters applied at this level will apply to the whole alert. 

2. Event Controller. Here' s where you define the KPI/Metric you would like to be alerted on. Filters on the Event level only apply to the Event, so you can set benchmarks tied to different metrics or dimensions and create multiple events on different components of your ad stack. 

3. Benchmark Controller. The benchmark is the standard that would be used as reference to evaluate fluctuations on the event. Filters applied on the Benchmark level only apply to the benchmarks.  

4. Trigger Controller. This is the value upon which the alert will be triggered. The triggered refers to any fluctuation on the metric set at the Event Controller level when being compared to the Benchmark. 

5. Schedule Controller. Allows you to set a duration for the alert to be active. Also allows you to appoint email addresses of the people you would like to be informed about incidentes occurring to your alert. 

6. Alerts Viewer. Here you can see the Status of all your created alerts and the number of incidents that have triggered the alert. . 

Bonus: You can Add multiple Events and Benchmarks by clicking on the Yellow button labeled as "Add Event". 

Definitions. 

1. Event. The metric or KPI that will be monitored by the alert  (e.g. Daily Revenue 300x250, CPM for Rubicon, etc.) 

2. Benchmark. The metric or KPI that will be used as reference to monitor the Event (e.g. Avg. Rev. last 7 days  300x250, Avg. CPM  all SSPs except Rubicon, etc).   

3. Trigger: the threshold or concrete fluctuation upon which the alert will be triggered. (e.g. 5% Decrease, 10% Increase, $0,30, etc). 

4. Incident: When the Event meets the criteria set on the Trigger with respecto of the Benchmark. 

5. Function: Within the context of setting Events, Benchmarks and Triggers, shall mean "Average", "Sum", "Constant", "%", "Increase" or "Decrease". 

6. Filter: Allows you to select a Metric or Dimension for setting up Events and Benchmarks. 

7. Metric: Any data point reported by your ad monetization partner (e.g. Revenue, CPM, Page Views, etc). 

8. Dimension: Any monetization partner, part of your stack, or layer applied to a metric (e.g. Ad Unit, Country, Connection/SSP, Size, Key Value, Ad Format, Inventory Type, Website). 

9. Occurrence: The number of times required for an event to happen before triggering an alert. 

10. Today: Means to the latest data reported today (normally SSPs report data from the previous day, however some SSPs report real time data). 

11. Yesterday: Means the data from the previous day to Today's data. 

12. Active Alert: An alert actively reporting incidents. 

13. Inactive Alert. An alert that has been paused and is no longer reporting on incidents. 

Adding and editing alerts.

1. Click on the edit icon.

2. Edit you alert on the alert generator.

Note: Alerts can be edited at all time, but results from previous alert settings will remain on the Incidents dashboard. 

3. To Delete an alert click on the Icon Delete (Garbage Can).

Note: You can only delete alerts before incidents are recorded for this alert. If an incident is recorded you can only Deactivate the alert.  

4. To Deactivate an alert simply change the Date to any point in the past from the Alert Generator. 

5. Don't forget to click on the Save "green button" to save your alerts. 

Alert modes. 

Generally speaking, there are 3 modes in which you can use the Alert system:

1. Surveillance Mode (Basic): Use the alerts feature to create simple alerts on KPIs that you need to monitor on daily basis. 

2. Discovery mode (Intermediate): Use the alerts feature to discover anomalies that are hard to catch with the naked eye. 

3. Experiments mode (Advanced): Use the alerts to run experiments to: 

  • Cross-tech experiments: Measure the impact of a new technology partner on your existing stack;

  • Cross-dimensional: Measure the impact across different layers of your stack  (e.g. impact of latency on revenue, impact of refresh on UX/Traffic, impact of Paywall revenue on Header Bidding, etc). 

  • Auction Dynamics: mix the different components of programmatic auctions into hybrid experiments to build winning strategies and massive auction advantages.

Use Cases. 

Use Case 1 (Surveillance Mode) - Spot sudden CPM drop in one of your top 3 supply sources. 

Problem statement: Your top SSPs amount for 55% of your Ad Revenue (Google AdExchange 30%, Index Exchange 18% and AppNexus 7%). Your overall WoW revenue shows an increase of 0.9% compared to the previous week and therefore there are no red flags popping up on your Revenue performance at first glance. Nevertheless a substantial CPM drop experienced by Index Exchange on the last 4 consecutive days of the week, combined with AppNexus increasing its fill rate 3x in one single day, made it hard to identify that money was left on the table by Index Exchange since the overall weekly revenue showed a small improvement.  After further investigation the AdOps team found out that the currency of the Rate on the Prebid related Line Item for Index Exchange was accidentally changed from US Dollars to HK Dollars thus competing on Price Priority at significantly lower rates. 

Impact: Shall Index Exchange CPM had remained stable during the 4 day negative streak, the Publisher should have experienced a much larger revenue increase that week. That would have required the AdOps team to monitor CPM fluctuations for all Supply sources on daily basis including weekends. 

Solution: Set individual alerts for each of your SSPs informing you when the daily CPM across all your supply sources decrease more than 15%-20% for at least 2 consecutive days compared to the avg. CPM experienced by such sources in the previous 7 days. 

you will always keep every fluctuation under the radar, even when such fluctuation doesn't necessary result on a decrease of the Revenue WoW growth rate or if that fluctuation occurs during weekends, bank holidays or leaves. 

Set up:  

Alert Name: SSP Name CPM drop in X%. 

Filter on main controller: Connection>Pick the desired SSP. 

Event 1: 

Filter on Event controller: None

Metric: CPM

Function: Avg. 

Range: Today. 

Benchmark 1

Filter on Benchmark controller: None

Metric: CPM

Function: Avg. 

Range: Last 7 days. 

Trigger: 

Trigger: Decrease

FunctIon: %

Value: 20%

Occurrence: 2 times

Schedule: Pick desired date.

Use Case 2 - (Discovery Mode). Impact of AMP delivery on Traffic. 

Problem statement: AMP reduces data size up to 8X making mobile sites run faster and thus increasing traffic and improving loading time. Despite of the pros, AMP also has some cons including higher operational workload and limited access to demand as not all ad networks support AMP or at least have less demand for AMP ad inventory. Strictly analyzing Revenue and Traffic fluctuations alone won't provide enough elements to attribute any fluctuation to AMP as seasonality and other factors could be involved. ​

Impact AMP might not be having the impact you expected on traffic and site speed, at least not to justify the trade off for getting lower non-AMP related Fill rates and ad revenue. 

Solution: Set an alert to inform you when "Time on Page" or "Session Duration" daily results, increase in over 10% when also AMP related Page Views or Ad Impressions increase in a similar 10% ratio. 

Set up:  

Alert Name: 10%+ Time on Page WHEN 10%+ AMP Page views.  

Filter on main controller: Inventory Type = AMP

Event 1: 

Filter on Event controller: Connection>Google Analytics

Metric: Session Duration/Time on Page

Function: Avg. 

Range: Today. 

Benchmark 1

Filter on Benchmark controller: Connection = Google Analytics

Metric: Session Duration/Time on Page

Function: Avg. 

Range: Last 7 days. 

Trigger: 

Trigger: Increase

FunctIon: %

Value: 10%

Occurrence: 1 time

Add Event

Event 2: 

Filter on Event controller: Connection = Google Analytics

Metric: Page Views or Ad Impressions

Function: Avg. 

Range: Today. 

Benchmark 1

Filter on Benchmark controller: Connection = Google Analytics

Metric: Page Views or Ad Impressions

Function: Avg. 

Range: Last 7 days. 

Trigger: 

Trigger: Increase

FunctIon: %

Value: 10%

Occurrence: 1 time

Schedule: Pick desired date.

Use Case 3 - (Experiment mode> Auction dynamics). - Bid Rate drops despite decrease on Timeout Rate. 

Problem statement: Your total Programmatic ad revenue from yesterday dropped 20% across the board. This could be caused by certain bidders being down or experiencing technical issues. In the past, the AdOps team identified that their Top SSP was experiencing a Time out rate 2 times higher than the rest of the SSPs in the mix and therefore the team decided to increase the universal timeout setup to give all bidders more room to breathe. Despite visible improvements of the Timeout rate across all SSPs, the Bid Rate of the Pubisher's top SSP (in other words the number of bids received from that SSP) kept dropping consistently, resulting on further revenue loss. The main challenge for the AdOps team is to troubleshoot the further causes for Bid Rate drop by looking into further metrics such as bid load time, avg. bid CPM, avg. Win CPM, win rate, etc., as this can vary from SSP to SSP and would require 24/7 surveillance.  

Impact: Eventhough the company has invested on a solid Prebid Analytics technology the amount of work and attention to detail required to spot every anomaly on the fly, would be extremely heavy making it almost impossible for an AdOps to keep every deviation under the radar. 

 

Solution: Set individual Alerts for each Header Bidding Partner covering all Prebid Analytics scenarios involving a decrease of the "Daily Win Rate"  in X% when: 

a. Time out drops in X%

b. Bid Response Time drops in X%

c. Bid Rate drops in X%

Set up Scenario 1: Win Rate drops WHEN Time out drops. 

Alert Name:10%+ Win Rate drop WHEN 10%+ Time out drop.  

Filter on main controller: Connection = SSP1

Event 1: 

Filter on Event controller: None

Metric: Win Rate

Function: Avg. 

Range: Today. 

Benchmark 1

Filter on Benchmark controller: None

Metric: Win Rate

Function: Avg. 

Range: Last 7 days. 

Trigger: 

Trigger: Increase

FunctIon: %

Value: 10%

Occurrence: 1 time

Add Event

Event 2: 

Filter on Event controller: Connection = SSP1

Metric: Time Out Rate

Function: Avg. 

Range: Today. 

Benchmark 1

Filter on Benchmark controller: Connection = SSP1

Metric: Time Out Rate

Function: Avg. 

Range: Last 7 days. 

Trigger: 

Trigger: Decrease

FunctIon: %

Value: 10%

Occurrence: 1 time

Schedule: Pick desired date.

For more examples or guidance on how to set your alerts visit our Slack channel or contact us at adops@yieldpass.com