{"id":"https://openalex.org/W7172005422","doi":"https://doi.org/10.1145/3770854.3783922","title":"Case Study: Learning Robust, Long-run Surrogate Metrics with Modeling and Instrumental Variables","display_name":"Case Study: Learning Robust, Long-run Surrogate Metrics with Modeling and Instrumental Variables","publication_year":2026,"publication_date":"2026-04-20","ids":{"openalex":"https://openalex.org/W7172005422","doi":"https://doi.org/10.1145/3770854.3783922"},"language":null,"primary_location":{"id":"doi:10.1145/3770854.3783922","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3770854.3783922","pdf_url":null,"source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 32nd ACM SIGKDD Conference on Knowledge Discovery and Data Mining V.1","raw_type":"proceedings-article"},"type":"conference-paper","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://doi.org/10.1145/3770854.3783922","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5144163340","display_name":"Daniel Day","orcid":"https://orcid.org/0009-0006-8875-6487"},"institutions":[{"id":"https://openalex.org/I4210099336","display_name":"Menlo School","ror":"https://ror.org/01240pn49","country_code":"US","type":"education","lineage":["https://openalex.org/I4210099336"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Daniel Day","raw_affiliation_strings":["Meta, Menlo Park, CA, USA"],"raw_orcid":"https://orcid.org/0009-0006-8875-6487","affiliations":[{"raw_affiliation_string":"Meta, Menlo Park, CA, USA","institution_ids":["https://openalex.org/I4210099336"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5144155603","display_name":"Matthew Dalton","orcid":"https://orcid.org/0009-0004-6347-3267"},"institutions":[{"id":"https://openalex.org/I4210099336","display_name":"Menlo School","ror":"https://ror.org/01240pn49","country_code":"US","type":"education","lineage":["https://openalex.org/I4210099336"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Matthew Dalton","raw_affiliation_strings":["Meta, Menlo Park, CA, USA"],"raw_orcid":"https://orcid.org/0009-0004-6347-3267","affiliations":[{"raw_affiliation_string":"Meta, Menlo Park, CA, USA","institution_ids":["https://openalex.org/I4210099336"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5075229217","display_name":"Samuel Daulton","orcid":"https://orcid.org/0000-0002-1448-7396"},"institutions":[{"id":"https://openalex.org/I4210099336","display_name":"Menlo School","ror":"https://ror.org/01240pn49","country_code":"US","type":"education","lineage":["https://openalex.org/I4210099336"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Samuel Daulton","raw_affiliation_strings":["Meta, Menlo Park, CA, USA"],"raw_orcid":"https://orcid.org/0000-0002-1448-7396","affiliations":[{"raw_affiliation_string":"Meta, Menlo Park, CA, USA","institution_ids":["https://openalex.org/I4210099336"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5043715747","display_name":"Benjamin Letham","orcid":"https://orcid.org/0009-0003-5205-8212"},"institutions":[{"id":"https://openalex.org/I4210099336","display_name":"Menlo School","ror":"https://ror.org/01240pn49","country_code":"US","type":"education","lineage":["https://openalex.org/I4210099336"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Benjamin Letham","raw_affiliation_strings":["Meta, Menlo Park, CA, USA"],"raw_orcid":"https://orcid.org/0009-0003-5205-8212","affiliations":[{"raw_affiliation_string":"Meta, Menlo Park, CA, USA","institution_ids":["https://openalex.org/I4210099336"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5144164429","display_name":"Xiao Cai","orcid":"https://orcid.org/0000-0002-6127-7068"},"institutions":[{"id":"https://openalex.org/I4210099336","display_name":"Menlo School","ror":"https://ror.org/01240pn49","country_code":"US","type":"education","lineage":["https://openalex.org/I4210099336"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Xiao Cai","raw_affiliation_strings":["Meta, Menlo Park, CA, USA"],"raw_orcid":"https://orcid.org/0000-0002-6127-7068","affiliations":[{"raw_affiliation_string":"Meta, Menlo Park, CA, USA","institution_ids":["https://openalex.org/I4210099336"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5144168331","display_name":"Aojia Zhao","orcid":"https://orcid.org/0009-0003-1133-5316"},"institutions":[{"id":"https://openalex.org/I4210099336","display_name":"Menlo School","ror":"https://ror.org/01240pn49","country_code":"US","type":"education","lineage":["https://openalex.org/I4210099336"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Aojia Zhao","raw_affiliation_strings":["Meta, Menlo Park, CA, USA"],"raw_orcid":"https://orcid.org/0009-0003-1133-5316","affiliations":[{"raw_affiliation_string":"Meta, Menlo Park, CA, USA","institution_ids":["https://openalex.org/I4210099336"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5006700143","display_name":"Eytan Bakshy","orcid":null},"institutions":[{"id":"https://openalex.org/I4210099336","display_name":"Menlo School","ror":"https://ror.org/01240pn49","country_code":"US","type":"education","lineage":["https://openalex.org/I4210099336"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Eytan Bakshy","raw_affiliation_strings":["Meta, Menlo Park, CA, USA"],"raw_orcid":"https://orcid.org/0009-0007-6480-9879","affiliations":[{"raw_affiliation_string":"Meta, Menlo Park, CA, USA","institution_ids":["https://openalex.org/I4210099336"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I4210099336"],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":null,"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"2165","last_page":"2175"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":null,"topics":[],"keywords":[{"id":"https://openalex.org/keywords/surrogate-model","display_name":"Surrogate model","score":0.7422999739646912},{"id":"https://openalex.org/keywords/metric","display_name":"Metric (unit)","score":0.673799991607666},{"id":"https://openalex.org/keywords/surrogate-endpoint","display_name":"Surrogate endpoint","score":0.5735999941825867},{"id":"https://openalex.org/keywords/surrogate-data","display_name":"Surrogate data","score":0.5577999949455261},{"id":"https://openalex.org/keywords/heuristics","display_name":"Heuristics","score":0.5073000192642212},{"id":"https://openalex.org/keywords/performance-metric","display_name":"Performance metric","score":0.45410001277923584},{"id":"https://openalex.org/keywords/gold-standard","display_name":"Gold standard (test)","score":0.43130001425743103},{"id":"https://openalex.org/keywords/outcome","display_name":"Outcome (game theory)","score":0.3725000023841858},{"id":"https://openalex.org/keywords/baseline","display_name":"Baseline (sea)","score":0.3343000113964081}],"concepts":[{"id":"https://openalex.org/C131675550","wikidata":"https://www.wikidata.org/wiki/Q7646884","display_name":"Surrogate model","level":2,"score":0.7422999739646912},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.6951000094413757},{"id":"https://openalex.org/C176217482","wikidata":"https://www.wikidata.org/wiki/Q860554","display_name":"Metric (unit)","level":2,"score":0.673799991607666},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6373000144958496},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.6212999820709229},{"id":"https://openalex.org/C141341695","wikidata":"https://www.wikidata.org/wiki/Q2368661","display_name":"Surrogate endpoint","level":2,"score":0.5735999941825867},{"id":"https://openalex.org/C142806159","wikidata":"https://www.wikidata.org/wiki/Q7646876","display_name":"Surrogate data","level":3,"score":0.5577999949455261},{"id":"https://openalex.org/C127705205","wikidata":"https://www.wikidata.org/wiki/Q5748245","display_name":"Heuristics","level":2,"score":0.5073000192642212},{"id":"https://openalex.org/C2780898871","wikidata":"https://www.wikidata.org/wiki/Q860554","display_name":"Performance metric","level":2,"score":0.45410001277923584},{"id":"https://openalex.org/C40993552","wikidata":"https://www.wikidata.org/wiki/Q514654","display_name":"Gold standard (test)","level":2,"score":0.43130001425743103},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.4207000136375427},{"id":"https://openalex.org/C148220186","wikidata":"https://www.wikidata.org/wiki/Q7111912","display_name":"Outcome (game theory)","level":2,"score":0.3725000023841858},{"id":"https://openalex.org/C12725497","wikidata":"https://www.wikidata.org/wiki/Q810247","display_name":"Baseline (sea)","level":2,"score":0.3343000113964081},{"id":"https://openalex.org/C5274069","wikidata":"https://www.wikidata.org/wiki/Q2285707","display_name":"Categorical variable","level":2,"score":0.3334999978542328},{"id":"https://openalex.org/C83546350","wikidata":"https://www.wikidata.org/wiki/Q1139051","display_name":"Regression","level":2,"score":0.3257000148296356},{"id":"https://openalex.org/C45804977","wikidata":"https://www.wikidata.org/wiki/Q7239673","display_name":"Predictive modelling","level":2,"score":0.31299999356269836},{"id":"https://openalex.org/C105795698","wikidata":"https://www.wikidata.org/wiki/Q12483","display_name":"Statistics","level":1,"score":0.3027999997138977},{"id":"https://openalex.org/C152877465","wikidata":"https://www.wikidata.org/wiki/Q208042","display_name":"Regression analysis","level":2,"score":0.30160000920295715},{"id":"https://openalex.org/C182365436","wikidata":"https://www.wikidata.org/wiki/Q50701","display_name":"Variable (mathematics)","level":2,"score":0.2994000017642975},{"id":"https://openalex.org/C117220453","wikidata":"https://www.wikidata.org/wiki/Q5172842","display_name":"Correlation","level":2,"score":0.2892000079154968},{"id":"https://openalex.org/C84525736","wikidata":"https://www.wikidata.org/wiki/Q831366","display_name":"Decision tree","level":2,"score":0.28839999437332153},{"id":"https://openalex.org/C48921125","wikidata":"https://www.wikidata.org/wiki/Q10861030","display_name":"Linear regression","level":2,"score":0.27459999918937683},{"id":"https://openalex.org/C43214815","wikidata":"https://www.wikidata.org/wiki/Q7310987","display_name":"Reliability (semiconductor)","level":3,"score":0.2700999975204468},{"id":"https://openalex.org/C148483581","wikidata":"https://www.wikidata.org/wiki/Q446488","display_name":"Feature selection","level":2,"score":0.2632000148296356},{"id":"https://openalex.org/C61797465","wikidata":"https://www.wikidata.org/wiki/Q1188986","display_name":"Term (time)","level":2,"score":0.257999986410141},{"id":"https://openalex.org/C111335779","wikidata":"https://www.wikidata.org/wiki/Q3454686","display_name":"Reduction (mathematics)","level":2,"score":0.25780001282691956},{"id":"https://openalex.org/C151956035","wikidata":"https://www.wikidata.org/wiki/Q1132755","display_name":"Logistic regression","level":2,"score":0.2529999911785126}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1145/3770854.3783922","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3770854.3783922","pdf_url":null,"source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 32nd ACM SIGKDD Conference on Knowledge Discovery and Data Mining V.1","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.1145/3770854.3783922","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3770854.3783922","pdf_url":null,"source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 32nd ACM SIGKDD Conference on Knowledge Discovery and Data Mining V.1","raw_type":"proceedings-article"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":13,"referenced_works":["https://openalex.org/W2112508839","https://openalex.org/W2155949904","https://openalex.org/W2575883117","https://openalex.org/W2809111384","https://openalex.org/W2986211311","https://openalex.org/W3124213571","https://openalex.org/W3124658416","https://openalex.org/W3133591098","https://openalex.org/W4386107890","https://openalex.org/W4401857172","https://openalex.org/W4401864323","https://openalex.org/W4409157819","https://openalex.org/W7133233387"],"related_works":[],"abstract_inverted_index":{"When":[0],"developing":[1],"product":[2,181],"changes":[3,284],"to":[4,15,38,113,172,303],"improve":[5],"long-run":[6,31,36,60,96,146,189,226,287],"user":[7],"outcomes,":[8,32],"surrogate":[9,55,68,79,128,141,161,184,199,216,244,281,305],"metrics":[10,56,69,80,129,142,217,245,267,282,306],"are":[11],"a":[12,41,121,170,235],"powerful":[13],"tool":[14],"help":[16],"accelerate":[17,48],"decision":[18,49],"making.":[19],"Randomized,":[20],"online":[21],"A/B":[22,88,176,212,229,253,274],"testing":[23],"has":[24],"become":[25],"the":[26,59,95,179,207,211,225,286],"gold":[27],"standard":[28],"for":[29,35,143,206,265,279],"measuring":[30],"but":[33,307],"waiting":[34],"outcomes":[37,190],"materialize":[39],"in":[40,72,81,152,174,178,238,271,285,316],"test":[42,213],"greatly":[43],"reduces":[44],"iteration":[45],"speed.":[46],"To":[47],"making,":[50],"we":[51,125,260,269],"instead":[52],"can":[53,311],"use":[54,173],"that":[57,164,276,293],"predict":[58],"outcome":[61,227],"more":[62,191,248],"quickly.":[63],"However,":[64,98],"finding":[65,77],"accurate,":[66],"robust":[67],"is":[70],"challenging":[71],"practice.":[73,317],"Existing":[74],"methods":[75,100],"propose":[76],"candidate":[78],"various":[82],"ways,":[83],"such":[84,104,163],"as":[85,105,169],"re-analyzing":[86],"past":[87],"experiments":[89],"where":[90,124,268],"candidates":[91],"strongly":[92],"correlate":[93],"with":[94,130,224],"outcome.":[97,289],"these":[99],"face":[101],"practical":[102],"limitations":[103],"being":[106],"affected":[107],"by":[108],"hidden":[109],"confounders":[110],"or":[111],"needing":[112],"rely":[114],"on":[115],"outdated,":[116],"historical":[117],"data.":[118],"We":[119,139,158],"developed":[120],"new":[122],"approach":[123],"learned":[126],"user-level":[127],"non-parametric":[131],"machine":[132,295],"learning":[133,296],"and":[134,150,154,231,297],"instrumental":[135],"variable":[136],"(IV)":[137],"analysis.":[138],"trained":[140],"two":[144],"key,":[145],"business":[147],"metrics,":[148],"BV1":[149,198],"BV2,":[151],"classification":[153],"regression":[155],"settings,":[156],"respectively.":[157],"deployed":[159],"each":[160,187],"model":[162],"it":[165],"would":[166],"be":[167],"available":[168],"metric":[171],"all":[175],"tests":[177,230,275],"target":[180,288],"surface.":[182],"Our":[183,243,290],"models":[185],"predicted":[186],"user's":[188],"accurately":[192],"than":[193,220],"baseline":[194,208],"heuristics":[195],"(e.g.,":[196],"our":[197,215,266,280],"model's":[200],"AUC":[201],"was":[202],"0.93":[203],"vs.":[204],"0.79":[205],"heuristic).":[209],"At":[210],"level,":[214],"showed":[218],"greater":[219],"0.9":[221],"Pearson":[222],"correlation":[223],"across":[228],"achieved":[232],"at":[233],"least":[234],"50%":[236],"improvement":[237],"accuracy":[239],"over":[240],"other":[241],"baselines.":[242],"were":[246],">30x":[247],"predictive":[249],"of":[250,263],"whether":[251],"an":[252],"test's":[254],"had":[255],"consistent":[256],"post-launch":[257],"effects.":[258],"Finally,":[259],"provide":[261],"evidence":[262,309],"surrogacy":[264],"show":[270],"multiple":[272],"independent":[273],"directly":[277],"optimizing":[278],"drove":[283],"results":[291],"demonstrate":[292],"combining":[294],"causal":[298],"inference":[299],"leads":[300],"not":[301],"only":[302],"accurate":[304],"also":[308],"they":[310],"act":[312],"like":[313],"true":[314],"surrogates":[315]},"counts_by_year":[],"updated_date":"2026-08-01T09:00:35.917206","created_date":"2026-08-01T00:00:00"}