This document discusses the end-to-end experimentation platform at GetYourGuide for A/B testing. It outlines the challenges of running experiments such as imbalanced assignments, suspicious metric changes, and non-converging results. It also describes the tools used for planning experiments, monitoring assignments, performing daily checks, and analyzing results. The goal is to validate UX changes, estimate effects on customers, and make more objective decisions through A/B testing while addressing issues that could impact experiment quality.
Using Cluster Analysis for Characteristics Detection in Software Defect ReportsIosif Itkin
Anna Gromova and Olga Moskaleva, Exactpro Researchers, participated in the Sixth International Conference on Analysis of Images, Social Networks and Texts, AIST 2017
AIST is a conference on the Analysis of Images, Social Networks, and Texts. This year, it was held at Moscow Polytechnic University on June 27-29. The conference is intended for computer scientists and industry professionals whose research interests involve applications of data mining and machine learning techniques to various problem domains: image processing, analysis of social networks, and natural language processing. AIST boasts its interdisciplinary nature, bridging the gap between researchers and practitioners.
All research papers submitted to the conference went through a double-blind reviewing process and were reviewed by at least three program committee members. The accepted papers were presented by their authors on July 27 and 28.
Anna Gromova’s research on Using Cluster Analysis for Characteristics Detection in Software Defect Reports was among this year’s accepted papers. Anna, who is an Exactpro researcher, presented her work in the General Topics of Data Analysis section of the conference.
Anna’s research focuses on the metrics and predictions of defect reports. In her research paper, Anna explores the workings of how extended bug attribute analysis can effectively reveal software weaknesses and improve the testing strategy, and how machine learning techniques can facilitate the process of building an automated recommendation system for project managers and QA team leads. Such a system would help them get a better picture of the risks associated with software defects, as well as obtain more accurate information about a bug's lifecycle.
“In future, we are planning to build an automated recommendation system based on the cluster analysis of our defect reports. A system like that could be used by all QA team members to improve the existing processes of developing the testing strategies and plans”, says Anna Gromova, Exactpro Researcher.
Using Cluster Analysis for Characteristics Detection in Software Defect ReportsIosif Itkin
Anna Gromova and Olga Moskaleva, Exactpro Researchers, participated in the Sixth International Conference on Analysis of Images, Social Networks and Texts, AIST 2017
AIST is a conference on the Analysis of Images, Social Networks, and Texts. This year, it was held at Moscow Polytechnic University on June 27-29. The conference is intended for computer scientists and industry professionals whose research interests involve applications of data mining and machine learning techniques to various problem domains: image processing, analysis of social networks, and natural language processing. AIST boasts its interdisciplinary nature, bridging the gap between researchers and practitioners.
All research papers submitted to the conference went through a double-blind reviewing process and were reviewed by at least three program committee members. The accepted papers were presented by their authors on July 27 and 28.
Anna Gromova’s research on Using Cluster Analysis for Characteristics Detection in Software Defect Reports was among this year’s accepted papers. Anna, who is an Exactpro researcher, presented her work in the General Topics of Data Analysis section of the conference.
Anna’s research focuses on the metrics and predictions of defect reports. In her research paper, Anna explores the workings of how extended bug attribute analysis can effectively reveal software weaknesses and improve the testing strategy, and how machine learning techniques can facilitate the process of building an automated recommendation system for project managers and QA team leads. Such a system would help them get a better picture of the risks associated with software defects, as well as obtain more accurate information about a bug's lifecycle.
“In future, we are planning to build an automated recommendation system based on the cluster analysis of our defect reports. A system like that could be used by all QA team members to improve the existing processes of developing the testing strategies and plans”, says Anna Gromova, Exactpro Researcher.
Here is the Six sigma training material part 7, presented by Skillogic Knowledge Solutions. If you are looking for Lean + Six Sigma training in Bangalore / Bengaluru, Visit Skillogic and schedule you classroom training.
How do you know if you have too much process, too little, or just the right amount? If you ignore process completely, unpredictability and chaos can follow. If you define the process to the nth degree and follow it religiously, the work grinds to a halt. Janet Gregory shares her experiences about how to find the tastiest balance of process and creativity for your projects and organization. She proposes that a formally defined process is sometimes necessary, but that it should be the exception. Explore with Janet the many variables—team size, complexity, criticality, organization structure, and culture—you must assess to find just the right balance. Learn how to make existing processes better by adding visibility to the process, getting team members’ input, and adapting documentation you need. Learn how to transform complicated processes into simpler ones—such as reporting a simple “thumbs up” or “thumbs down”—and go home with new tools to sprinkle on just enough process.
Choosing the right QA strategy for a successful projectThe Software House
Dawid Krala, Tomasz Górski: When talking about Quality in software development, the most important thing is to understand the huge long-term impact of sticking (or not) to your QA strategy. That’s why The Software House’s most experienced QA engineers are going to show it on real-life examples from our projects – going over the most important trends, concepts and good practices.
What will you learn from this presentation:
When is QA the most important factor behind successful software projects?
How to develop quality software and not to go bankrupt?
How to choose the most optimal testing tools for your business?
Which types of tests will be the best for your software project?
What are the practical differences between Quality Assurance and Quality Control?
How can you use the Test Pyramid in 2020?
What is the “extended standard approach” in software testing (integration tests, component tests, contract tests, E2E, static code analysis)?
This webinar aims to show you how to avoid the common pitfalls that can affect your conversion rate optimization strategy. At the end of this webinar you will be able to:
1) Use A/B Testing best practices to get the most out of it
2) Identify the common mistakes preemptively
3) Appreciate how A/B Testing fits your whole conversion rate optimization strategy (based on real cases, such as Swedoffice and Spotify)
hosted by Simon Dahla from Conversionista Sweden and Martijn Janssen, Partner Manager Optimizely
Narendra Ponnuswamy - Performance Testing Effort - Estimation or Guesstimation? TEST Huddle
EuroSTAR Software Testing Conference 2009 presentation on Performance Testing Effort - Estimation or Guesstimation? by Narendra Ponnuswamy. See more at conferences.eurostarsoftwaretesting.com/past-presentations/
Agile Development of High Performance ApplicationsFabian Lange
Slides from my talk at gearconf 2010 in Düsseldorf, discussing Performance as an important non-functional requirement. Because NFRs are hard to test, I showed how AppDynamics Lite could be used to ease pain and build better performing apps.
If you are interested in performance and application performance monitoring, visit our blog:
http://blog.codecentric.de/en/category/performance-en/
If you want to try appdynamics lite yourself, download it at http://appdynamics.com/free
Risk based monitoring presentation into what Risk Based Monitoring (RBM) is all about. The value of getting it right, and the risks of getting it wrong. We look at the Triumph Research Intelligence (TRI) approach to RBM, and the TRI solution to RBM. An insight into the Future of Visual OPRA is given.
Data Natives Frankfurt v 11.0 | "Competitive advantages with knowledge graphs...Dataconomy Media
The challenges of increasing complexity of organizations, companies and projects are obvious and omnipresent. Everywhere there are connections and dependencies that are often not adequately managed or not considered at all because of a lack of technology or expertise to uncover and leverage the relationships in data and information. In his presentation, Axel Morgner talks about graph technology and knowledge graphs as indispensable building blocks for successful companies.
More Related Content
Similar to Data NativesBerlin v 20.0 | "Serving A/B experimentation platform end-to-end" - Eugene Klyuchnikov
Here is the Six sigma training material part 7, presented by Skillogic Knowledge Solutions. If you are looking for Lean + Six Sigma training in Bangalore / Bengaluru, Visit Skillogic and schedule you classroom training.
How do you know if you have too much process, too little, or just the right amount? If you ignore process completely, unpredictability and chaos can follow. If you define the process to the nth degree and follow it religiously, the work grinds to a halt. Janet Gregory shares her experiences about how to find the tastiest balance of process and creativity for your projects and organization. She proposes that a formally defined process is sometimes necessary, but that it should be the exception. Explore with Janet the many variables—team size, complexity, criticality, organization structure, and culture—you must assess to find just the right balance. Learn how to make existing processes better by adding visibility to the process, getting team members’ input, and adapting documentation you need. Learn how to transform complicated processes into simpler ones—such as reporting a simple “thumbs up” or “thumbs down”—and go home with new tools to sprinkle on just enough process.
Choosing the right QA strategy for a successful projectThe Software House
Dawid Krala, Tomasz Górski: When talking about Quality in software development, the most important thing is to understand the huge long-term impact of sticking (or not) to your QA strategy. That’s why The Software House’s most experienced QA engineers are going to show it on real-life examples from our projects – going over the most important trends, concepts and good practices.
What will you learn from this presentation:
When is QA the most important factor behind successful software projects?
How to develop quality software and not to go bankrupt?
How to choose the most optimal testing tools for your business?
Which types of tests will be the best for your software project?
What are the practical differences between Quality Assurance and Quality Control?
How can you use the Test Pyramid in 2020?
What is the “extended standard approach” in software testing (integration tests, component tests, contract tests, E2E, static code analysis)?
This webinar aims to show you how to avoid the common pitfalls that can affect your conversion rate optimization strategy. At the end of this webinar you will be able to:
1) Use A/B Testing best practices to get the most out of it
2) Identify the common mistakes preemptively
3) Appreciate how A/B Testing fits your whole conversion rate optimization strategy (based on real cases, such as Swedoffice and Spotify)
hosted by Simon Dahla from Conversionista Sweden and Martijn Janssen, Partner Manager Optimizely
Narendra Ponnuswamy - Performance Testing Effort - Estimation or Guesstimation? TEST Huddle
EuroSTAR Software Testing Conference 2009 presentation on Performance Testing Effort - Estimation or Guesstimation? by Narendra Ponnuswamy. See more at conferences.eurostarsoftwaretesting.com/past-presentations/
Agile Development of High Performance ApplicationsFabian Lange
Slides from my talk at gearconf 2010 in Düsseldorf, discussing Performance as an important non-functional requirement. Because NFRs are hard to test, I showed how AppDynamics Lite could be used to ease pain and build better performing apps.
If you are interested in performance and application performance monitoring, visit our blog:
http://blog.codecentric.de/en/category/performance-en/
If you want to try appdynamics lite yourself, download it at http://appdynamics.com/free
Risk based monitoring presentation into what Risk Based Monitoring (RBM) is all about. The value of getting it right, and the risks of getting it wrong. We look at the Triumph Research Intelligence (TRI) approach to RBM, and the TRI solution to RBM. An insight into the Future of Visual OPRA is given.
Similar to Data NativesBerlin v 20.0 | "Serving A/B experimentation platform end-to-end" - Eugene Klyuchnikov (20)
Data Natives Frankfurt v 11.0 | "Competitive advantages with knowledge graphs...Dataconomy Media
The challenges of increasing complexity of organizations, companies and projects are obvious and omnipresent. Everywhere there are connections and dependencies that are often not adequately managed or not considered at all because of a lack of technology or expertise to uncover and leverage the relationships in data and information. In his presentation, Axel Morgner talks about graph technology and knowledge graphs as indispensable building blocks for successful companies.
Data Natives Munich v 12.0 | "How to be more productive with Autonomous Data ...Dataconomy Media
Every day we are challenged with more data, more use cases and an ever increasing demand for analytics. In this talk Bjorn will explain how autonomous data management and machine learning help innovators to more productive and give examples how to deliver new data driven projects with less risk at lower costs.
Data Natives meets DataRobot | "Build and deploy an anti-money laundering mo...Dataconomy Media
Compliance departments within banks and other financial institutions are turning to machine learning for improving their Anti Money Laundering compliance activities. Today, the systems that aim to detect potentially suspicious activity are commonly rule-based, and suffer from ultra-high false positive rates. DataRobot will discuss how their Automated Machine Learning platform was successfully used for a real use case to reduce their false positives and to enhance their Anti-Money Laundering activities.
Data Natives Munich v 12.0 | "Political Data Science: A tale of Fake News, So...Dataconomy Media
Trump, Brexit, Cambridge Analytica... In the last few years, we have had to confront the consequences of the use and misuse of data science algorithms in manipulating public opinion through social media. The use of private data to microtarget individuals is a daily practice (and a trillion-dollar industry), which has serious side-effects when the selling product is your political ideology. How can we cope with this new scenario?
Data Natives Vienna v 7.0 | "The Ingredients of Data Innovation" - Robbert de...Dataconomy Media
When taking a deep dive into the world of data, one thing is certain: the ultimate goal is to create something new, something better, something faster. In other words, innovation should always be at the forefront of companies strategic outlook, whether their goal is to pioneer new processes, user experiences, products or services.
Data Natives Cologne v 4.0 | "The Data Lorax: Planting the Seeds of Fairness...Dataconomy Media
What does it take to build a good data product or service? Data practitioners always think about the technology, user experience and commercial viability. But rarely do they think about the implications of the systems they build. This talk will shed light on the impact of AI systems and the unintended consequences of the use of data in different products. It will also discuss our role, as data practitioners, in planting the seeds of fairness in the systems we build.
Data Natives Cologne v 4.0 | "How People Analytics Can Reveal the Hidden Aspe...Dataconomy Media
We all hear about the power of data, big data and data analysis in todays market place. But rarely feel it's touchable effects on our own business decisions and performance.
Let's dive into it and see how can people analytics increase people performance, motivation and business revenue?
Data Natives Amsterdam v 9.0 | "Ten Little Servers: A Story of no Downtime" -...Dataconomy Media
Cloud Infrastructure is a hostile environment: a power supply failure or a network outage leads to downtime and big losses. There is nothing we can trust: a single server, a server rack, even a whole datacenter can fail, and if an application is fragile by design, disruption is inevitable. We must distribute our application and diversify cloud data strategy to survive disturbances of any scale. Apache Cassandra is a cloud-native platform-agnostic database that stores data with a distributed redundancy so it easily survives any issue. What to know how Apple and Netflix handle petabytes of data, keeping it highly available? Join us and listen to a story of 10 little servers and no downtime!
Data Natives Amsterdam v 9.0 | "Point in Time Labeling at Scale" - Timothy Th...Dataconomy Media
In the data industry, having correctly labelled datasets is vital. Timothy Thatcher explains how tagging your data while considering time and location and complex hierarchical rules at scale can be handled.
Data Natives Berlin v 20.0 | "Ten Little Servers: A Story of no Downtime" - A...Dataconomy Media
Cloud Infrastructure is a hostile environment: a power supply failure or a network outage leads to downtime and big losses. There is nothing we can trust: a single server, a server rack, even a whole datacenter can fail, and if an application is fragile by design, disruption is inevitable. We must distribute our application and diversify cloud data strategy to survive disturbances of any scale. Apache Cassandra is a cloud-native platform-agnostic database that stores data with a distributed redundancy so it easily survives any issue. What to know how Apple and Netflix handle petabytes of data, keeping it highly available? Join us and listen to a story of 10 little servers and no downtime!
Big Data Frankfurt meets Thinkport | "The Cloud as a Driver of Innovation" - ...Dataconomy Media
Creativity is the mental ability to create new ideas and designs. Innovation, on the other hand, Means developing useful solutions from new ideas. Creativity can be goal-oriented, Whereas innovation is always goal-oriented. This bedeutet, dass innovation aims to achieve defined goals. The use of cloud services and technologies promises enterprise users many benefits in terms of more flexible use of IT resources and faster access to innovative solutions. That’s why we want to examine the question in this talk, of what role cloud computing plays for innovation in companies.
Thinkport meets Frankfurt | "Financial Time Series Analysis using Wavelets" -...Dataconomy Media
Presentation of Time Series Properties of Financial Instrument and Possibilities in Frequency Decomposition and Information Extraction using FT, STFT and Wavelets with Outlook in Current Research on Wavelet Neural Networks
Big Data Helsinki v 3 | "Distributed Machine and Deep Learning at Scale with ...Dataconomy Media
"With most machine learning (ML) and deep learning (DL) frameworks, it can take hours to move data for ETL, and hours to train models. It's also hard to scale, with data sets increasingly being larger than the capacity of any single server. The amount of the data also makes it hard to incrementally test and retrain models in near real-time.
Learn how Apache Ignite and GridGain help to address limitations like ETL costs, scaling issues and Time-To-Market for the new models and help achieve near-real-time, continuous learning.
Yuriy Babak, the head of ML/DL framework development at GridGain and Apache Ignite committer, will explain how ML/DL work with Apache Ignite, and how to get started.
Topics include:
— Overview of distributed ML/DL including architecture, implementation, usage patterns, pros and cons
— Overview of Apache Ignite ML/DL, including built-in ML/DL algorithms, and how to implement your own
— Model inference with Apache Ignite, including how to train models with other libraries, like Apache Spark, and deploy them in Ignite
— How Apache Ignite and TensorFlow can be used together to build distributed DL model training and inference"
Big Data Helsinki v 3 | "Federated Learning and Privacy-preserving AI" - Oguz...Dataconomy Media
"Machine learning algorithms require significant amounts of training data which has been centralized on one machine or in a datacenter so far. For numerous applications, such need of collecting data can be extremely privacy-invasive. Recent advancements in AI research approach this issue by a new paradigm of training AI models, i.e., Federated Learning.
In federated learning, edge devices (phones, computers, cars etc.) collaboratively learn a shared AI model while keeping all the training data on device, decoupling the ability to do machine learning from the need to store the data in the cloud. From personal data perspective, this paradigm enables a way of training a model on the device without directly inspecting users’ data on a server. This talk will pinpoint several examples of AI applications benefiting from federated learning and the likely future of privacy-aware systems."
Big Data Helsinki v 3 | "What you should know about PSD2 APIs?" - Joonas TomperiDataconomy Media
PSD2 is coming live in September 2019 and after that, you are able to get account data, make strong authentication and make payments through banks PSD2 APIs.
Let's dive deeper into the world of ODC! Ricardo Alves (OutSystems) will join us to tell all about the new Data Fabric. After that, Sezen de Bruijn (OutSystems) will get into the details on how to best design a sturdy architecture within ODC.
Neuro-symbolic is not enough, we need neuro-*semantic*Frank van Harmelen
Neuro-symbolic (NeSy) AI is on the rise. However, simply machine learning on just any symbolic structure is not sufficient to really harvest the gains of NeSy. These will only be gained when the symbolic structures have an actual semantics. I give an operational definition of semantics as “predictable inference”.
All of this illustrated with link prediction over knowledge graphs, but the argument is general.
Software Delivery At the Speed of AI: Inflectra Invests In AI-Powered QualityInflectra
In this insightful webinar, Inflectra explores how artificial intelligence (AI) is transforming software development and testing. Discover how AI-powered tools are revolutionizing every stage of the software development lifecycle (SDLC), from design and prototyping to testing, deployment, and monitoring.
Learn about:
• The Future of Testing: How AI is shifting testing towards verification, analysis, and higher-level skills, while reducing repetitive tasks.
• Test Automation: How AI-powered test case generation, optimization, and self-healing tests are making testing more efficient and effective.
• Visual Testing: Explore the emerging capabilities of AI in visual testing and how it's set to revolutionize UI verification.
• Inflectra's AI Solutions: See demonstrations of Inflectra's cutting-edge AI tools like the ChatGPT plugin and Azure Open AI platform, designed to streamline your testing process.
Whether you're a developer, tester, or QA professional, this webinar will give you valuable insights into how AI is shaping the future of software delivery.
Essentials of Automations: Optimizing FME Workflows with ParametersSafe Software
Are you looking to streamline your workflows and boost your projects’ efficiency? Do you find yourself searching for ways to add flexibility and control over your FME workflows? If so, you’re in the right place.
Join us for an insightful dive into the world of FME parameters, a critical element in optimizing workflow efficiency. This webinar marks the beginning of our three-part “Essentials of Automation” series. This first webinar is designed to equip you with the knowledge and skills to utilize parameters effectively: enhancing the flexibility, maintainability, and user control of your FME projects.
Here’s what you’ll gain:
- Essentials of FME Parameters: Understand the pivotal role of parameters, including Reader/Writer, Transformer, User, and FME Flow categories. Discover how they are the key to unlocking automation and optimization within your workflows.
- Practical Applications in FME Form: Delve into key user parameter types including choice, connections, and file URLs. Allow users to control how a workflow runs, making your workflows more reusable. Learn to import values and deliver the best user experience for your workflows while enhancing accuracy.
- Optimization Strategies in FME Flow: Explore the creation and strategic deployment of parameters in FME Flow, including the use of deployment and geometry parameters, to maximize workflow efficiency.
- Pro Tips for Success: Gain insights on parameterizing connections and leveraging new features like Conditional Visibility for clarity and simplicity.
We’ll wrap up with a glimpse into future webinars, followed by a Q&A session to address your specific questions surrounding this topic.
Don’t miss this opportunity to elevate your FME expertise and drive your projects to new heights of efficiency.
Search and Society: Reimagining Information Access for Radical FuturesBhaskar Mitra
The field of Information retrieval (IR) is currently undergoing a transformative shift, at least partly due to the emerging applications of generative AI to information access. In this talk, we will deliberate on the sociotechnical implications of generative AI for information access. We will argue that there is both a critical necessity and an exciting opportunity for the IR community to re-center our research agendas on societal needs while dismantling the artificial separation between the work on fairness, accountability, transparency, and ethics in IR and the rest of IR research. Instead of adopting a reactionary strategy of trying to mitigate potential social harms from emerging technologies, the community should aim to proactively set the research agenda for the kinds of systems we should build inspired by diverse explicitly stated sociotechnical imaginaries. The sociotechnical imaginaries that underpin the design and development of information access technologies needs to be explicitly articulated, and we need to develop theories of change in context of these diverse perspectives. Our guiding future imaginaries must be informed by other academic fields, such as democratic theory and critical theory, and should be co-developed with social science scholars, legal scholars, civil rights and social justice activists, and artists, among others.
Key Trends Shaping the Future of Infrastructure.pdfCheryl Hung
Keynote at DIGIT West Expo, Glasgow on 29 May 2024.
Cheryl Hung, ochery.com
Sr Director, Infrastructure Ecosystem, Arm.
The key trends across hardware, cloud and open-source; exploring how these areas are likely to mature and develop over the short and long-term, and then considering how organisations can position themselves to adapt and thrive.
Smart TV Buyer Insights Survey 2024 by 91mobiles.pdf91mobiles
91mobiles recently conducted a Smart TV Buyer Insights Survey in which we asked over 3,000 respondents about the TV they own, aspects they look at on a new TV, and their TV buying preferences.
Epistemic Interaction - tuning interfaces to provide information for AI supportAlan Dix
Paper presented at SYNERGY workshop at AVI 2024, Genoa, Italy. 3rd June 2024
https://alandix.com/academic/papers/synergy2024-epistemic/
As machine learning integrates deeper into human-computer interactions, the concept of epistemic interaction emerges, aiming to refine these interactions to enhance system adaptability. This approach encourages minor, intentional adjustments in user behaviour to enrich the data available for system learning. This paper introduces epistemic interaction within the context of human-system communication, illustrating how deliberate interaction design can improve system understanding and adaptation. Through concrete examples, we demonstrate the potential of epistemic interaction to significantly advance human-computer interaction by leveraging intuitive human communication strategies to inform system design and functionality, offering a novel pathway for enriching user-system engagements.
Epistemic Interaction - tuning interfaces to provide information for AI support
Data NativesBerlin v 20.0 | "Serving A/B experimentation platform end-to-end" - Eugene Klyuchnikov
1. Building the future
of experiential travel
Johannes Reck
Serving A/B experimentation
platform end-to-end
Eugene Klyuchnikov
2. Europe’s largest marketplace
for travel experiences
50k+
Products in 150+
countries
25M+
Tickets sold
$650M+
In VC funding
600+
Strong global team
150+
Traveler nationalities
3. We make it simple to book and enjoy
incredible experiences
4. 4
To validate UX changes
To estimate the effect
To understand what our
customers like
To be more objective
Because we can!
Because correlation
is not causation
Why to run A/B tests?
5. 5
Correlation is not causation
- seasonality?
- marketing
effect?
- random
fluctuations?
random factors
and 3rd party
effects are
eliminated
6. 6
Architecture
Applications
Application B
.
.
.
Application Z
“Raw” events Enriched events
Application A
enrichment job
- filters out office IPs,
- filters out bots and crawlers,
- detects suspicious behavior,
- etc.
A/B experiments
summary
experiment summary job
- calculates all relevant metrics for all
active experiments
- performs cumulative summarization
7. 7
Challenge #1
Applications
● Events are not being sent
● Wrong events are sent
● Events miss some critical information
● Completely imbalanced assignment
due to technical issues
Application B
.
.
.
Application Z
“Raw” events Enriched events
Application A
A/B experiments
summary
Early
monitoring
8. 8
Challenge #1
● Kibana
● Don’t care about slight imbalance
● Near real-time monitoring
● All environments
● Immediate feedback for developers
9. 9
Challenge #2
Applications
● Imbalanced behavior (too many bots,
redirects, etc. on one variation or user group)
● Unreasonably low / high number of visitors
● Suspicious behavior
● Bizarre funnels
Application B
.
.
.
Application Z
“Raw” events Enriched events
Application A
A/B experiments
summary
Experiment
planning
Early
analysis
10. 10
Challenge #2
● Looker + common sense
● Number of visitors should match the plan
● Share of total visitors should be stable
● Sometimes cohort analysis
11. 11
Challenge #3
Applications
● Statistically imbalanced assignments
(sometimes small)
● Non-converging / suspicious uplifts
● Significant changes in the funnel
● Money burn
Application B
.
.
.
Application Z
“Raw” events Enriched events
Application A
A/B experiments
summary
Daily
monitoring
Automatic
alerts
12. 12
Challenge #3
● Historical uplift (convergence)
● Assignment balance (chi-sq. test)
● Switchers below the threshold
● Money impact is acceptable
● Guardrail metrics feel good
13. 13
Grey area #1
Grey Area #1
● event naming conventions
● event firing conventions
● timing conventions
● event containers
● on- / off-boarding events
● etc. etc.
● Defining the standards
● Regular syncs
● Training
● Documentation
14. 14
Grey area #2
Grey Area #2
● rules for stopping experiment
● interpreting the results
● understanding funnel impact
● multidirectional metrics
● multiple comparisons problem
(the dead salmon syndrome)
● etc. etc.
● Defining the standards
● Regular syncs
● Training
● Documentation
15. 15
Experiment tooling from end to end
Plan
experiment
Sample size tool
● Estimate the
duration of an
experiment
● Understand the
impact of limiting
to certain
segments on run
time
Dig deeper
Experiment funnel
analysis
● Configure a
funnel and see if
an experiment
had a significant
impact on any of
the steps
● Explore from here
to add more filters
on funnel steps
Get a team
overview
Team experiment
overview
● See all currently
active trials per
team and their
impact
● Estimate the
overall
test-over-test
contribution of
experiments to CR
Analyze results
Experiment dashboard
● See the impact of
an experiment on
success and
support metric
● See the remaining
run time till uplift
detection
Monitor
assignment
Kibana dashboard
● See the number of
events in near real
time
● Check the
assignment
balance between
A and B
● Estimate the
duration of an
experiment
● Understand the
impact of limiting
to certain
segments on run
time
● See the number of
events in near real
time
● Check the
assignment
balance between
A and B