Test Driven Relevancy, Presented by Doug Turnbull at SolrExchage DC
Download as PPTX, PDF0 likes1,030 views
Doug discusses challenges with collaboration between search developers and content experts when optimizing search relevancy. The current process of developers making changes and experts having to wait a week for results is inefficient. Doug proposes applying test-driven development principles to search by having experts continuously test search results and provide feedback on changes in real-time. This allows developers to get immediate feedback and ensures changes are improving search quality. Doug's company built a tool called Quepid that implements this approach to enable better collaboration between experts and developers when optimizing search.
Test Driven Relevancy, Presented by Doug Turnbull at SolrExchage DC
- 2. Test Driven Relevancy Doug Turnbull Search Relevancy Expert, OpenSource Connections
- 3. Hi I’m Doug! Its me! Also introducing my non-present client Rena… @softwaredoug http://www.linkedin.com/in/softwaredoug http://bit.ly/softwaredoug renam@silverchair.com OpenSource Connections
- 4. How do sales/content curators collaborate with devs? “Myocardial in-what-tion? Dangit Rena, I’m a Solr consultant – not a doctor” “When doctors search for ‘myocardial infarction’ these documents about ‘heart attacks’ should come up” “I don’t even…” “let me work my Solr magic and get back to you next week…” OpenSource Connections
- 5. How do content curators collaborate with devs? • Doug knows Solr q={!boost b=log(numCatPictures)}bunnies • Rena knows her content Myocardial infarction Renal carcinoma<tokenizer class="solr.WhitespaceTokenizerFactory"/> • Bob knows his business Sales Niche customers Conversions Supplier Pressure This is a universal pattern – it takes different strokes! OpenSource Connections
- 6. How do content curators collaborate with devs? “Rena, I fixed that myocardial in-whatever-tion relevancy issue” “Ok but you broke this other thing I thought was fixed!” <reiterates that he’s a search expert not a doctor> <reiterates she’s a paying client> “ok let me see what I can do…. I’ll get back to you in a week” OpenSource Connections
- 7. • People Problem: Our collaboration stinks o “Throw problems over the fence… wait a week” o Siloed by expertise (search vs content experts) o Potential for politics, anger, frustration (all paths to the dark side) • Technical Problem: Search testing is hard o Small set of relevancy rules impacts all user searches o Therefore: Much easier to have regressions than in software o Very easy to have regressions Our Problems OpenSource Connections
- 8. • In short, Broken Workflow: o Iterations between devs and non-technical experts take a long time o Devs need immediate access to non-technical expertise to make rapid progress • Gather broken searches • Rate existing searches • Find searches that have slid backwards in quality o Non-technical experts clearly need devs • Translate business rules to relevancy rules • Bending sophisticated token matching engine to our approximate user intent Our Problems OpenSource Connections
- 9. • Our lack of collaboration means our testing stinks o Need expert feedback to test Our Problems • Search devs often don’t know good search! They need help. “I need an army of Renas locked in a room telling me what is good and bad” “this is good, this is bad, this is ok….” OpenSource Connections
- 10. • In s/w development -- automated testing is often away to collaborate o Devs Can sit together with content experts and ask: • What should happen in this case? o Then record that in the form of a test Solutions? @Given("tab $asciiTab") public void tab(String asciiTab) { tab = new Tab(asciiTab, new TabParser()); } @When("the guitar plays") public void guitarPlays() { guitar = new Guitar(); guitar.play(tab); } @Then("the following notes will be played $notes") public void theFollowingNotesWillBePlayed(String notes) { ensureThat(expectedNotes(notes), arePlayedBy(guitar)); } OpenSource Connections
- 11. • Collaborative testing is absolutely essential for search o Good search is best defined by experts in the content: • Marketing, sales, users, etc Test Driven Development with Search I’m a search expert! Not a Doctor! How can I possibly measure search Help me help you! I have few ways to record, measure, and evaluate search • Unfortunately there’s nothing to help content experts communicate with search devs around search (frankly this is rather shocking to me) OpenSource Connections
- 12. • Collaborative testing is absolutely essential for search o Every change to search relevancy will cause other queries to change o MUST know how much our relevancy has changed Test Driven Development with Search “I fixed your search… Does it matter that toe fungus query changed by 30%?” “Yeah lets see if we can work together to balance the two relevancy concerns” “I’m glad we have tests to know what’s changed!” OpenSource Connections
- 13. • Apply these ideas to search quality: o Given query Y o What documents are awesome/good/ok/bad/terrible? • Record these ratings somewhere • Observe the changes of all queries simultaneously as we modify relevancy params Test Driven Development with Search Now I’ve got the ultimate relevancy workbench. I can see if my ideas are working or failing right away! Now I can see instantly if Doug’s changes are making progress or if we’re moving backwards! OpenSource Connections
- 14. • We built a tool (a product!) around these ideas • Now our favorite relevancy workbench Quepid! “Hey we’re kicking butt and taking names on these search queries!” http://quepid.com Try Quepid! OpenSource Connections
- 15. Search Quality is about Collaboration* *and collaboration is about testing Test Driven Development with Quepid Demo time OpenSource Connections