CHAPTER 11 · APPENDIX
附录：术语、方法与引用
术语、方法与引用来源。引用索引已整理为可检索的独立页面。
附录 A：引用来源汇总#
全文共引用 370 个唯一链接（按出现顺序）：
370 条引用已整理为可检索索引
打开引用索引 →
按章节分组、支持关键词与域名筛选，并保留每条的来源标题与域名。
展开原始 URL 清单（370 条）
- https://www.reuters.com/legal/transactional/openai-acquire-product-testing-startup-statsig-appoints-cto-applications-2025-09-02/
- https://www.bloomberg.com/news/articles/2025-09-02/openai-to-buy-product-testing-startup-statsig-for-1-1-billion
- https://support.google.com/analytics/answer/12979939?hl=en
- https://monetate.com/resource/how-does-the-sunsetting-of-oracle-maxymiser-impact-cx-and-personalization/
- https://www.harness.io/press-and-news/harness-completes-acquisition-of-split-software
- https://www.businesswire.com/news/home/20250630514541/en/Monetate-Acquires-SiteSpect-to-Deliver-AI-Native-Personalization-and-Testing-at-Enterprise-Scale
- https://wingify.com/news/vwo-and-ab-tasty-unite-as-wingify/
- https://en.yna.co.kr/view/RPR20260916007800353?section=press-release/index
- https://www.tmcnet.com/usubmit/2026/07/30/10422711.htm
- https://www.sitecore.com/resources/insights/artificial-intelligence/cms-and-conversion-optimization
- https://www.optimizely.com/company/press/episerver-completes-acquisition-of-optimizely
- https://www.centrelanepartners.com/2022/10/28/kibo-spins-out-personalization-business-under-the-monetate-brand/
- https://www.sitecore.com/company/newsroom/press-releases/2021/03/sitecore-acquires-boxever-and-four51
- https://www.dbta.com/Editorial/News-Flashes/Oracle-Snaps-Up-Customer-Experience-Optimization-Specialist-to-Expand-Marketing-Cloud-106113.aspx
- https://www.optimizely.com/company/press/optimizely-acquires-netspring
- https://www.optimizely.com/de/company/press/opal-workforce
- https://certification.optimizely.com/docs/guides/experimentation/configure-mutual-exclusion/
- https://certification.optimizely.com/docs/concepts/experimentation/performance-edge/
- https://certification.optimizely.com/docs/concepts/experimentation/stats-engine-deep-dive/
- https://support.optimizely.com/hc/en-us/articles/33424529987597-CUPED-Controlled-experiment-Using-Pre-Experiment-Data
- https://support.optimizely.com/hc/en-us/articles/13409080412173-Optimizely-s-automatic-sample-ratio-mismatch-detection
- https://support.optimizely.com/hc/en-us/articles/42734792850061-Configure-a-Bayesian-A-B-test
- https://www.optimizely.com/plans/
- https://www.vendr.com/marketplace/optimizely
- https://help.wingify.com/hc/en-us/articles/58766108516889-Configure-Holdout-Groups-in-Feature-Experimentation
- https://wingify.com/pricing/
- https://help.wingify.com/hc/en-us/articles/58831743266969-About-Wingify-Data360
- https://wingify.com/blog/new-stats-engine-and-enhanced-wingify-reports/
- https://help.wingify.com/hc/en-us/articles/59066074450841-What-is-the-Math-Behind-SmartStats
- https://help.wingify.com/hc/en-us/articles/58838252139545-Achieve-Accurate-Campaign-Results-with-SmartStats-Configuration
- https://docs.abtasty.com/web-experimentation-and-personalization/campaign-creation-and-dashboard/types-of-campaigns
- https://docs.abtasty.com/account/tag-integration/understand-ab-tasty-tag-implementation-methods/choose-between-synchronous-and-asynchronous-methods
- https://docs.abtasty.com/server-side/sdks/sdk-overview
- https://docs.abtasty.com/reporting-and-performances/reporting/frequentist-analysis-mode
- https://docs.abtasty.com/help-center/faq/faq-sequential-testing-alerts
- https://docs.abtasty.com/reporting-and-performances/reporting/sample-ratio-mismatch
- https://docs.kameleoon.com/developer-docs/web-experimentation/technical-concepts/flicker-management-and-performance
- https://docs.kameleoon.com/user-manual/experiment-analytics/statistical-methods/choosing-the-right-statistical-method-for-ab-testing
- https://docs.kameleoon.com/user-manual/experiment-analytics/analyze-results/results-page-tabs/health-menu
- https://www.einpresswire.com/article_print/852526197/kameleoon-launches-vibe-experimentation-with-prompt-based-testing
- https://support.convert.com/hc/en-us/articles/360015870371-server-side-testing-
- https://support.convert.com/hc/en-us/articles/statistical-models-in-convertcoms-ab-testing-platform
- https://www.convert.com/pricing/
- https://doc.sitespect.com/knowledge/sitespect-implementation
- https://monetate.com/
- https://doc.sitespect.com/knowledge/sitespect-and-single-page-applications
- https://doc.sitespect.com/knowledge/sitespect-statistics
- https://experienceleague.adobe.com/en/docs/target/using/activities/activities
- https://www.miaprova.com/blog/two-customers-asked-us-the-same-question-inside-of-a-week-what-exactly-do-we-get-if-we-move-from-adobe-target-standard-to-premium/
- https://experienceleague.adobe.com/en/docs/target/using/activities/auto-allocate/automated-traffic-allocation
- https://experienceleague.adobe.com/en/docs/target/using/activities/auto-target/auto-target-to-optimize
- https://experienceleague.adobe.com/en/docs/target/using/activities/automated-personalization/algo-random-forest
- https://experienceleague.adobe.com/en/docs/target-dev/developer/server-side/on-device-decisioning/overview
- https://www.miaprova.com/blog/adobe-target-had-a-quietly-big-year-at-summit/
- https://business.adobe.com/products/target/pricing.html
- https://kirro.io/adobe-target-ab-testing
- https://experienceleague.adobe.com/zh-hant/docs/core-services/interface/data-collection/rdc?lang=zh-Hant
- https://www.cnbc.com/amp/2019/03/25/mcdonalds-announces-acquisition-to-personalize-the-drive-thru.html
- https://newsroom.mastercard.com/news/europe/de-de/newsroom/pressemitteilungen/de-de/2022/april/mastercard-starkt-mit-abschluss-der-ubernahme-von-dynamic-yield-seine-services-zur-kundenbindung/
- https://www.mastercard.com/us/en/business/consumer-acquisition-and-engagement/personalization/dynamic-yield.html
- https://support.dynamicyield.com/hc/en-us/articles/360034513414-Dynamic-Allocation
- https://dy.dev/docs/choose
- https://www.vendr.com/marketplace/dynamic-yield
- https://monetate.com/resource/monetate-introduces-forte/
- https://monetate.com/maestro/
- https://docs.monetate.com/docs/implementation-overview
- https://docs.monetate.com/docs/enable-auto-promotion-for-experiences
- https://docs.monetate.com/docs/create-an-automated-personalization-experience
- https://www.sitecore.com/platform/conversion-optimization
- https://developers.sitecore.com/learn/faq/xm-personalize/personalize-cdp
- https://doc.sitecore.com/personalize/en/users/sitecore-personalize/introduction-to-experiments-in-sitecore-personalize/introduction-to-experiments-in-sitecore-personalize.html
- https://doc.sitecore.com/personalize/en/users/sitecore-personalize/introduction-to-decisioning-in-sitecore-personalize.html
- https://doc.sitecore.com/personalize/en/users/sitecore-personalize/introduction-to-experiments-in-sitecore-personalize/running-an-experiment-with-optimized-testing/understanding-multi-armed-bandit-calculations.html
- https://www.vendr.com/marketplace/sitecore
- https://clonepartner.com/blog/sitecore-xp-end-of-support-2026-the-version-by-version-guide
- https://www.21cloudbox.com/support/sitecore-xm-cloud-china.html
- https://www.gartner.com/reviews/product/oracle-maxymiser-legacy
- https://origin.peerspot.com/products/comparisons/monetate-intelligent-personalization-engine_vs_oracle-maxymiser-eol
- https://developer.harness.io/docs/feature-management-experimentation/split-to-harness/migrate-from-split-to-harness/how-it-works
- https://statsig.com/blog/openai-acquisition
- https://docs.statsig.com/experiments/overview
- https://docs.statsig.com/experiments/holdouts-introduction
- https://docs.statsig.com/statsig-warehouse-native/features/statistics
- https://statsig.com/blog/announcing-cure
- https://docs.statsig.com/statsig-warehouse-native/native-vs-cloud
- https://docs.statsig.com/experiments/monitor
- https://docs.statsig.com/statsig-warehouse-native/features/experiment-options
- https://amplitude.com/blog/amplitude-and-statsig-partnership
- https://statsig.com/blog/statsig-amplitude-phase-1
- https://statsig.com/pricing
- https://docs.geteppo.com/reference/architecture-and-security/
- https://docs.geteppo.com/statistics/cuped/
- https://docs.geteppo.com/statistics/multiple-testing/
- https://docs.geteppo.com/data-management/organizing-metrics/guardrails
- https://docs.geteppo.com/statistics/global-lift/
- https://docs.geteppo.com/statistics/interaction-detection/
- https://www.datadoghq.com/about/latest-news/press-releases/datadog-acquires-eppo-to-expand-its-ai/
- https://docs.growthbook.io/overview
- https://github.com/growthbook/growthbook
- https://github.com/growthbook/growthbook/blob/main/LICENSE
- https://docs.growthbook.io/statistics/overview
- https://docs.growthbook.io/statistics/sequential
- https://docs.growthbook.io/statistics/cuped
- https://www.growthbook.io/pricing
- https://www.growthbook.io/blog/how-to-migrate-from-statsig-to-growthbook
- https://amplitude.com/docs/data/warehouse-native/overview
- https://amplitude.com/docs/faq/sequential-testing
- https://amplitude.com/docs/feature-experiment/experiment-theory/bayesian-statistics
- https://amplitude.com/docs/feature-experiment/troubleshooting/sample-ratio-mismatch
- https://amplitude.com/docs/feature-experiment/implementation
- https://amplitude.com/docs/feature-experiment/local-evaluation
- https://amplitude.com/pricing
- https://leaderpost.com/press-releases/globe-newswire/launchdarkly-surpasses-200-million-in-arr-as-demand-for-runtime-control-accelerates/
- https://launchdarkly.com/docs/guides/statistical-methodology/choosing
- https://launchdarkly.com/docs/guides/statistical-methodology/cuped
- https://launchdarkly.com/blog/warehouse-native-experimentation-comes-to-bigquery-databricks-and-redshift/
- https://launchdarkly.com/how-it-works/platform-architecture/
- https://launchdarkly.com/blog/what-happened-what-we-learned-and-how-were-improving/
- https://launchdarkly.com/pricing/
- https://developer.harness.io/feature-management-experimentation/use-fme/experimentation/key-concepts/fixed-horizon
- https://developer.harness.io/docs/feature-management-experimentation/sdks-and-infrastructure/
- https://github.com/Flagsmith/flagsmith
- https://docs.flagsmith.com/experimentation
- https://www.flagsmith.com/pricing
- https://configcat.com/docs/advanced/variation-id-for-analytics
- https://configcat.com/pricing/
- https://www.getunleash.io/blog/unleash-moving-to-agplv3
- https://www.getunleash.io/blog/series-b
- https://support.google.com/analytics/answer/12979939
- https://firebase.google.com/docs/ab-testing/ab-concepts
- https://aws.amazon.com/blogs/mt/support-for-amazon-cloudwatch-evidently-ending-soon/
- https://aws.amazon.com/blogs/devops/running-production-experiments-with-aws-appconfig-experimentation/
- https://blog.cloudflare.com/flagship/
- https://www.volcengine.com/product/DataTester
- https://www.365trade.com.cn/jfwzb/775230.jhtml
- https://www.alibabacloud.com/help/zh/pai/a-b-experiment/
- http://www.gdebidding.com/nhpbgs/199863.jhtml
- https://www.cac.gov.cn/2024-03/22/c_1712776611775634.htm
- https://learn.microsoft.com/en-us/azure/azure-app-configuration/concept-feature-management
- https://openfeature.dev/
- https://aicon.infoq.cn/2025/shanghai/speaker/9508
- https://docs.byteplus.com/en/docs/DataTester/Learn-about-A-B-Testing-Products-DataTester
- https://docs.byteplus.com/en/docs/DataTester/Traffic-Shunt-Agent
- https://docs.byteplus.com/en/docs/DataTester/Basic-Concept
- https://www.doit.com.cn/p/475545.html
- https://docs.byteplus.com/en/docs/DataTester/Traffic-Layer-and-Mutual-Exclusion-Domain
- https://docs.byteplus.com/en/docs/DataTester/How-to-Interpret-the-Experiment-Report
- https://docs.byteplus.com/en/docs/DataTester/Enhanced-Shunt-Uniformity
- https://docs.byteplus.com/en/docs/DataTester/MAB-Multi-Armed-Bandit-Intelligent-Tuning-Experiment
- https://server.it168.com/a2024/1023/6866/000006866793.shtml
- https://arxiv.org/abs/2305.02542
- https://developer.aliyun.com/article/1672360
- https://zhuanlan.zhihu.com/p/390417035
- https://zhuanlan.zhihu.com/p/394253252
- https://arxiv.org/abs/2201.00689
- https://help.aliyun.com/zh/open-search/industry-algorithm-edition/a-or-b-tests
- https://help.aliyun.com/en/ahas/product-overview/delisting-announcement-of-function-switch-function-module
- https://cloud.tencent.com.cn/developer/article/2156664
- https://cloud.tencent.com.cn/developer/article/2346725
- https://arxiv.org/abs/2405.08411
- https://patents.google.com/patent/CN110830234B/zh
- https://docs.abetterchoice.ai/guide/experiments/stats-engine-and-validation
- https://arxiv.org/abs/2408.04441
- https://doi.org/10.1145/3711896.3736906
- https://tech.meituan.com/2024/09/09/New-Generation-Analysis-Engine-Driving-Data-Decisions.html
- https://zhuanlan.zhihu.com/p/147203168
- https://arxiv.org/abs/2202.10887
- https://ziyuan.baidu.com/wiki/2727
- https://cloud.baidu.com/doc/AI_REC/s/Flw4scq90
- https://www.selectdb.com/blog/1719
- https://arxiv.org/abs/2608.04625
- https://zhuanlan.zhihu.com/p/715370120
- https://github.com/bilibili/abtest
- https://www.bilibili.com/read/cv21106590/
- https://zhuanlan.zhihu.com/p/10465095421
- https://testerhome.com/column_channels/38235
- https://zhuanlan.zhihu.com/p/587995471
- https://zhuanlan.zhihu.com/p/641222646
- https://zhuanlan.zhihu.com/p/25685006
- https://cloud.tencent.cn/developer/article/2536676
- https://www.sohu.com/a/848315597_121665362
- https://www.infoq.cn/article/T7HCVlTeUCSroDRhT1FF
- https://news.e-works.net.cn/category802/news115463.htm
- https://docs.byteplus.com/en/docs/DataTester/Server-Side-SDK-Overall-Description
- https://www.infoq.cn/article/alimama-large-scale-online-hierarchical-experiment
- https://www.infoq.cn/article/tmall-app-ab-test/
- https://tech.meituan.com/2025/08/22/meituan-AB-Online-Controlled-Experiment-08.html
- https://tech.meituan.com/2019/11/28/advertising-performance-experiment-configuration-platform.html
- https://s3plus.meituan.net/ddfile/2025%E5%8F%AF%E4%BF%A1%E5%AE%9E%E9%AA%8C%E7%99%BD%E7%9A%AE%E4%B9%A62025.8.20.pdf
- https://arxiv.org/abs/2602.08569
- https://blog.itpub.net/70048664/viewspace-3184138/
- https://eureka.patsnap.com/patent/CN115391184B
- http://www.npc.gov.cn/npc/c2/c30834/202108/t20210820_313088.html
- https://www.gov.cn/zhengce/zhengceku/2022-01/04/content_5666429.htm
- https://www.growthbook.io/about
- https://docs.growthbook.io/warehouses
- https://docs.growthbook.io/self-host
- https://docs.growthbook.io/self-host/production
- https://docs.growthbook.io/faq
- https://docs.growthbook.io/lib/build-your-own
- https://docs.growthbook.io/statistics/details
- https://docs.growthbook.io/app/experiment-results
- https://docs.growthbook.io/statistics/multiple-corrections
- https://docs.growthbook.io/self-host/env
- https://posthog.com/pricing
- https://posthog.com/docs/experiments/creating-an-experiment
- https://posthog.com/docs/how-posthog-works
- https://posthog.com/docs/how-posthog-works/clickhouse
- https://github.com/PostHog/post-mortems/blob/main/2025-09-29-flags-is-down.md
- https://posthog.com/docs/experiments/statistics-bayesian
- https://posthog.com/docs/experiments/statistics-frequentist
- https://posthog.com/docs/experiments/cuped
- https://posthog.com/docs/experiments/exposures
- https://posthog.com/docs/feature-flags/cutting-costs
- https://posthog.com/docs/self-host
- https://github.com/Unleash/unleash/releases/tag/v8.0.0
- https://docs.getunleash.io/get-started/unleash-overview
- https://docs.getunleash.io/concepts/activation-strategies
- https://docs.getunleash.io/guides/a-b-testing
- https://docs.flagsmith.com/experimentation/analyse-an-experiment
- https://docs.flagsmith.com/experimentation/statistics
- https://docs.flagsmith.com/performance/edge-proxy
- https://docs.flipt.io/v2/concepts
- https://docs.flipt.io/v2/configuration/analytics
- https://docs.flipt.io/v2/licensing
- https://github.com/openflagr/flagr
- https://openflagr.github.io/flagr/flagr_overview
- https://openflagr.github.io/flagr/flagr_exposure
- https://openflagr.github.io/flagr/flagr_eval_exposure_pipeline
- https://openfeature.dev/specification/
- https://flagd.dev/
- https://openfeature.dev/specification/sections/tracking
- https://www.kdd.org/kdd2019/accepted-papers/view/diagnosing-sample-ratio-mismatch-in-online-controlled-experiments-a-taxonom
- https://github.com/FeatureProbe/FeatureProbe/blob/main/README_CN.md
- https://featureprobe.github.io/FeatureProbe/zh-CN/reference/structure/
- https://featureprobe.github.io/FeatureProbe/zh-CN/reference/evaluation-rules/
- https://featureprobe.github.io/FeatureProbe/zh-CN/reference/metric-explanation/
- https://featureprobe.github.io/FeatureProbe/blog/FeatureProbe%20bayesian/
- https://docs.abrouter.com/docs/managing/makingdecisions/
- https://superposition.juspay.io/docs/basic-concepts/context-aware-config/how-all-this-works/
- https://superposition.juspay.io/docs/basic-concepts/experimentation/experiment-groups/
- https://www.getunleash.io/pricing
- https://www.exp-platform.com/Documents/2013-02-CUPED-ImprovingSensitivityOfControlledExperiments.pdf
- https://arxiv.org/abs/2004.12489
- https://engineering.atspotify.com/2026/1/why-we-use-separate-tech-stacks-for-personalization-and-experimentation
- https://doi.org/10.1145/1835804.1835810
- https://www.itl.nist.gov/div898/handbook/pri/section3/pri333.htm
- https://arxiv.org/abs/2311.14698
- https://arxiv.org/abs/1501.00450
- https://docs.statsig.com/experiments/layers-overview
- https://docs.growthbook.io/app/holdouts
- https://docs.geteppo.com/experiment-analysis/holdouts/
- https://arxiv.org/abs/1903.08755
- https://arxiv.org/abs/2009.00148
- https://docs.geteppo.com/experiment-analysis/switchbacks/
- https://arxiv.org/abs/2508.00751
- https://arxiv.org/abs/2303.10094
- https://docs.geteppo.com/sdks/architecture/overview/
- https://raw.githubusercontent.com/optimizely/go-sdk/refs/tags/v0.1.0/optimizely/decision/experiment_bucketer_service.go
- https://docs.statsig.com/sdks/how-evaluation-works
- https://launchdarkly.github.io/ruby-server-sdk/LaunchDarkly/Impl/EvaluatorBucketing.html
- https://docs.geteppo.com/sdks/event-logging/assignment-logging/
- https://docs.geteppo.com/statistics/confidence-intervals/statistical-nitty-gritty/
- https://docs.growthbook.io/statistics/quantile
- https://docs.statsig.com/experiments/statistical-methods/methodologies/cuped
- https://www.evanmiller.org/how-not-to-run-an-ab-test.html
- https://www.optimizely.com/field-notes/articles/statistics-for-the-internet-age-the-story-behind-optimizelys-new-stats-engine
- https://docs.geteppo.com/statistics/confidence-intervals/analysis-methods/
- https://docs.geteppo.com/statistics/sample-ratio-mismatch/
- https://docs-legacy.statsig.com/experiments-plus/monitor/
- https://www.microsoft.com/en-us/research/publication/diagnosing-sample-ratio-mismatch-in-online-controlled-experiments-a-taxonomy-and-rules-of-thumb-for-practitioners/
- https://docs.growthbook.io/experimentation-analysis/cluster-experiments
- https://www.cambridge.org/core/books/abs/trustworthy-online-controlled-experiments/metrics-for-experimentation-and-the-overall-evaluation-criterion/0A8BF7F06A1338FF6D10C49F4CD8A198
- https://docs.geteppo.com/data-management/definitions/overview/
- https://docs.geteppo.com/data-management/metrics/
- https://docs.growthbook.io/app/metrics
- https://eu.36kr.com/zh/p/3730988847890694
- https://docs.geteppo.com/data-management/certified-metrics/
- https://docs.geteppo.com/data-management/data-pipeline/
- https://docs.geteppo.com/experiment-analysis/configuration/filter-assignments-by-entry-point/
- https://docs.geteppo.com/data-management/definitions/assignment-sql/
- https://blog.csdn.net/m0_60025795/article/details/129880151
- https://www.modb.pro/db/604087
- https://docs.geteppo.com/guides/advanced-experimentation/anonymous-explainer/
- https://docs.kameleoon.com/user-manual/experiment-analytics/analyze-results/data-and-metrics/how-kameleoon-counts-conversions#attribution-window
- https://www.lukasvermeer.nl/srm/docs/faq/
- https://www.lukasvermeer.nl/srm/docs/causes/
- https://druid.apache.org/docs/latest/design/
- https://docs.snowflake.com/en/release-notes/2025/other/2025-09-23-snowpipe-streaming-high-performance-architecture
- https://cloud.google.com/bigquery/docs/write-api-streaming
- https://docs.geteppo.com/data-management/warehouse-best-practices/
- https://docs-legacy.statsig.com/statsig-warehouse-native/guides/costs
- https://arxiv.org/abs/1910.03878
- https://research.google/pubs/overlapping-experiment-infrastructure-more-better-faster-experimentation/
- https://launchdarkly.com/docs/sdk/relay-proxy
- https://docs.getunleash.io/unleash-edge
- https://docs.growthbook.io/self-host/proxy
- https://launchdarkly.com/blog/designing-for-failure-to-avoid-disaster/
- https://openfeature.dev/specification/sections/flag-evaluation
- https://pkg.go.dev/github.com/launchdarkly/go-server-sdk/v7@v7.13.3/ldcomponents
- https://developer.harness.io/docs/feature-management-experimentation/sdks-and-infrastructure/optional-infra/split-daemon-splitd/
- https://docs.getunleash.io/get-started/unleash-overview.md
- https://openfeature.dev/specification/sections/hooks
- https://openfeature.dev/specification/appendix-c/
- https://firebase.google.com/docs/remote-config/use-cases/minimum-version
- https://docs.geteppo.com/feature-flagging/concepts/targeting/
- https://firebase.google.com/docs/ab-testing/ab-concepts?hl=en
- https://developer.apple.com/documentation/storekit/skadnetwork
- https://developers.cloudflare.com/flagship/reference/limits/
- https://docs.getunleash.io/deploy/hosting-options.md
- https://docs.getunleash.io/concepts/rbac.md
- https://docs.getunleash.io/concepts/change-requests.md
- https://launchdarkly.com/docs/home/releases/approvals
- https://docs.getunleash.io/concepts/events.md
- https://launchdarkly.com/docs/home/releases/version-restore
- https://martinfowler.com/articles/feature-toggles.html
- https://docs.getunleash.io/concepts/technical-debt.md
- https://news.bjd.com.cn/2023/08/29/10546096.shtml
- https://sre.google/workbook/implementing-slos/
- https://principlesofchaos.org/
- https://docs.aws.amazon.com/fis/latest/userguide/what-is.html
- https://tech.meituan.com/2025/05/22/meituan-AB-Online-Controlled-Experiment-01.html
- https://confidence.spotify.com/glossary/experimentation-maturity-model
- https://atticusli.com/blog/posts/how-netflix-runs-experimentation/
- https://research.chalmers.se/publication/507720
- https://atticusli.com/blog/posts/how-booking-com-runs-experimentation/
- https://research.google.com/pubs/archive/36500.pdf
- https://engineering.atspotify.com/2020/10/29/spotifys-new-experimentation-platform-part-1/
- https://confidence.spotify.com/docs/capabilities/experiments/reviews.md
- https://www.geteppo.com/blog/make-decisions-before-experimenting
- https://docs.geteppo.com/experiment-analysis/configuration/protocols/
- https://docs.growthbook.io/kb/experiments/aa-tests
- https://docs.statsig.com/experiments/monitor.md
- https://docs.geteppo.com/data-management/organizing-metrics/guardrails/
- https://docs-legacy.statsig.com/safeguards/overview/
- https://gdpr-info.eu/art-5-gdpr/
- https://oag.ca.gov/privacy/ccpa
- https://www.gov.cn/lianbo/bumen/202502/content_7003767.htm
- https://www.gov.cn/zhengce/2022-01/04/content_5728941.htm
- https://arxiv.org/abs/2407.10457
- https://arxiv.org/abs/2306.05685
- https://arxiv.org/abs/2606.17165
- https://netflixtechblog.com/sequential-a-b-testing-keeps-the-world-streaming-netflix-part-1-continuous-data-cba6c7ed49df
- https://medium.com/airbnb-engineering/https-medium-com-jonathan-parks-scaling-erf-23fd17c91166
- https://partner.booking.com/en-gb/click-magazine/industry-perspectives/role-experimentation-bookingcom
- https://www.microsoft.com/en-us/research/publication/the-anatomy-of-a-large-scale-experimentation-platform/
- https://dl.acm.org/doi/abs/10.1145/3292500.3330722
- https://experimentguide.com/
- https://research.google/pubs/pub36500/
- https://speakerdeck.com/srihari/overlapping-experim
- https://ar5iv.labs.arxiv.org/html/1409.3174
- https://www.sigops.org/s/conferences/sosp/2015/current/2015-Monterey/008-tang-online.pdf
- https://www.linkedin.com/blog/engineering/ab-testing-experimentation/making-the-linkedin-experimentation-engine-20x-faster
- https://dl.acm.org/doi/10.1145/2783258.2788602
- https://www.uber.com/blog/xp/
- https://engineering.atspotify.com/2025/9/spotifys-experiments-with-learning-framework
- https://engineering.atspotify.com/2026/9/why-spotify-is-not-using-bayesian-a-b-testing
- https://www.etsy.com/codeascraft/imbalance-detection-for-healthier-experimentation
- https://opendatascience.com/experimentation-platform-at-doordash/
- https://tech.instacart.com/variance-reduction-below-the-randomization-grain-31719f87a7d2
- https://tech.instacart.com/geo-based-a-b-testing-and-difference-in-difference-analysis-in-instacart-catalog-2ba05e744a3f
- https://www.amazon.science/publications/science-of-price-experimentation-at-amazon
- https://sellercentral.amazon.com/help/hub/reference/external/GYZYA77YZTWTF7YZ
- https://ar5iv.labs.arxiv.org/html/2405.08411
- https://blog.csdn.net/didi_tech/article/details/106424045
- https://www.cambridge.org/core/books/trustworthy-online-controlled-experiments/D97B26382EB0EB2DC2019A7A7B518F59
- https://dl.acm.org/doi/10.1145/2433396.2433413
- https://ar5iv.labs.arxiv.org/html/1512.04922
- https://www.optimizely.com/contentassets/9205a8a811e84957a7cca527d4af20be/whitepaper_optimizely_stats_engine.pdf/download
- https://www.datadoghq.com/blog/datadog-acquires-eppo/
附录 B：术语表#
- A/B/n：同时比较两个及以上变体的对照实验。
- CUPED：Controlled-experiment Using Pre-Experiment Data，用实验前数据做协变量以降低方差、提升灵敏度。
- SRM：Sample Ratio Mismatch，实际分流比例与设计比例显著不符，通常意味着实验实现或数据链路有 bug。
- OEC：Overall Evaluation Criterion，用于权衡短期与长期、收益与代价的总体评估指标。
- MDE：Minimum Detectable Effect，给定功效与样本量下能检测到的最小效应。
- Sequential Testing：允许在实验过程中多次查看并提前停止的统计方法，控制整体假阳性率。
- Delta Method：用泰勒展开近似非线性指标（如比率、人均值）方差的方法。
- Holdout：保留一部分用户长期不进入任何实验，用于衡量实验体系整体增量。
- Orthogonal Layers：用不同 salt/层让多个实验在同一批用户上正交分流，互不干扰。
- Switchback：按时间片交替切换实验组/对照组，适用于双边市场与强干扰场景。
- Interleaving：把多个候选结果交错混排，用偏好比较替代绝对指标，适合搜索/推荐排序。
- MAB：Multi-Armed Bandit，在探索与利用之间动态分配流量。
- Warehouse-Native：实验平台不存储明细数据，指标计算下推到客户数仓。
- OpenFeature：CNCF 的 Feature Flag 开放标准，用于解耦 SDK 与供应商。
- Guardrail Metrics：护栏指标，用于确保实验不损害体验、性能、收入等底线。
- Novelty Effect：新功能初期带来的短期新鲜感效应，可能导致实验高估长期收益。
附录 C：研究方法与分工#
- L0 基线：市场大事记、40 家厂商清单、14 个开源项目 GitHub 活跃度。
- 深度研究（25 路并发）：A1-A6 商业平台；B1-B5 中国平台；C1-C5 开源；D1-D6 技术模块；E1-E3 案例与学术；F1-F2 交叉验证与缺口补全。
- 分册合成（7 路）：S-A1、S-A2、S-B、S-C、S-D1、S-D2、S-E。
- 终稿：执行摘要、市场地图、选型矩阵、参考架构与路线图。
本报告由多个研究 agent 并行产出后人工/主控合成。所有价格、融资、收购信息以文中标注日期为准；采购前请以供应商最新报价与合同为准。
NOTE本页内容来自《AB 测试平台全网竞品与方案调研》2026-10 版。价格、融资与产品状态以文中标注日期为准；采购前请以供应商最新报价与合同为准。