Evaluation of Websites.md (2486B)
1 +++ 2 title = 'Evaluation of Websites' 3 +++ 4 # Evaluation of Websites 5 ## Why evaluate? 6 7 - attract more visitors 8 - sell more products 9 - decide which web app to use 10 - improve visitor ratings 11 - etc. 12 13 ## Evaluation studies 14 15 - always start with a clear research question 16 - aka "problem statement" 17 - practical/theoretical relevance, feasible 18 - types of studies 19 - explorative — what is related? 20 - e.g. why do people visit the site again? 21 - descriptive — what happens? 22 - e.g. how many people find the site through a SE? 23 - explanatory — why does it happen? 24 - if you add login, do people visit the site again? 25 - does a change in structure make it easier for visitors to find what they're looking for? 26 - make a hypothesis 27 - a prediction of outcome of test 28 - deduced from theory or observations 29 - collect data 30 - e.g. in lab experiments, survey, interview… 31 - qualitative (non-numerical) and quantitative (numerical) 32 - test dependent var with respect to independent var 33 - specific population (customers/all web users/registered users/whatever) 34 - specific sample (random/convenience/volunteers) 35 - evaluation methods 36 - common 37 - mockups 38 - low fidelity — early in design phase, only basic functionality, static, cheap. focus on concepts. 39 - high fidelity — later in design phase, refined details, expensive. 40 - prototypes — working example of website 41 - focus groups — moderated group discussion, early in design stage. 42 - card sorting — group of people sort items into clusters to get intuitive structure for website 43 - usability inspection — go systematically through website, check against Ten Web Guidelines. performed by dev team. 44 - group walkthrough — group of people walk through website as if performing primary tasks 45 - user testing — remote, observe user while primary tasks are performed. log actions, eye tracking, record video/audio. 46 - survey 47 48 - specific to web evaluation 49 - web analytics 50 - analyse logfiles 51 - JavaScript page tagging to capture visitor data 52 - very good objective data, but privacy concerns and no insight into motivation or unvisited pages. 53 - online experiments 54 - distribute visitors over versions, see which performs better 55 - after release 56 - example: A/B testing