Bezkoda vietņu veidotāju validācija pēc lietotāju atsauksmēm un popularitātes

Abstract

Patērētāju lēmumu pieņemšana ir pakļauta sociālajām ietekmēm (piemēram, lietotāju atsauksmes un popularitātes reitingi), kuras veicina pārliecību par produkta kvalitāti. Tiešsaistes popularitātes reitingi, no kuriem patērētāji ietekmējas, nereti sniedz neprecīzu popularitātes atainojumu, jo tie tiek balstīti tikai uz produktu piedāvātajām funkcionalitātēm un ir subjektīvi, jo neņem vērā lietotāju viedokli par produktu un mēdz būt finansiāli stimulēti. Šī darba mērķis ir noteikt, vai tiešsaistes popularitātes reitingi ir uzticami un vai attēlotā bezkoda vietņu veidotāju popularitāte tiešsaistes reitingos ir pielīdzināma tam, kā lietotāji vērtē šos rīkus. Eksistējošie pētījumi norāa, ka lietotāju atsauksmes ieņem nozīmīgu lomu produktu popularitātē un lietotāju apmierinātībā. Turklāt publikācijas vēsta, ka lietotāju atsauksmju analīze ar dabiskās valodas apstrādi spēj sniedz vērtīgu ieskatu lietotāju apmierinātībā un ļauj noteikt pieejamo funkciju nozīmīgumu. Šī pētījuma praktiskā daļa sastāv no bezkoda vietņu veidotāju tiešsaistes popularitātes reitingu izpētes, vietņu veidotāju lietotāju atsauksmju rasmošanas, lietotāju atsauksmju atslēgvārdu iegūšanas un rīku lietotāju skaita noteikšanas. Ievāktās atsauksmes tiek analizētas, izmantojot dabiskās valodas apstrādi, lai noteiktu lietotāju sentimentu un atsauksmēs minētās funkcijas. Iegūtie dati apstiprina, ka tiešsaistes popularitātes reitingu uzticamība ir apšaubāma. Neskatoties uz to, lietotāju viedoklis par veidotāju popularitāti sakrīt ar tiešsaistē atainoto popularitāti. Tālākiem pētījumiem vajadzētu veikt padziļinātu datu analīzi, kuras pamatā būtu datu apjoma palielināšana, kā arī atslēgvārdu un atsauksmju sentimenta sasaistīšana. Diplomdarba valoda ir angļu. Tas sastāv no 84 lappusēm, 12 attēliem, 4 tabulām un 4 pielikumiem, un tajā ir apskatīti 87 literatūras avoti. Atslēgvārdi: bezkoda vietņu veidotāji, lietotāju atsauksmes, popularitāte, sentimenta analīze, atslēgvārdu analīze

No-code website builders are tools for website development that provide users with pre-made website elements and an intuitive interface and do not require programming knowledge or experience. There are currently more than 50 website builders available on the market, and the continuous growth of the number of builders is mainly due to the increasing demand for more affordable website development options (WebFX, n.d.). Consumers care not only for product features but also for product recommendations (Pitman, 2022). While having a wide selection offer, consumers tend to make decisions based on collective judgment, reflected as product popularity (Hanson & Putler, 1996; Laland, 2004). To find out the popularity of the available tools, consumers consider online available popularity rankings. These rankings have a high probability of reflecting inaccurate popularity due to different incentives that benefit ranking creators and, at the same time, increase ranking subjectivity (Murton Beets & Handley, 2019). Additionally, preliminary observations show that most of the online found popularity rankings are only based on tool offered features despite the consumer care for user reviews and experience. Hence, this research is done to determine whether online popularity rankings are reliable and if the depicted builder popularity in online rankings is comparable to how users perceive the builders. Online popularity ranking reliability was determined by qualitative analysis of 20 nocode website builder popularity rankings. The qualitative data collected were observations of subjectivity reinforcing fact presence, and the builders ranked in the publications and their ranking order. The data confirmed the suspicion of online ranking subjectivity as 17 out of 20 rankings were monetarily incentivized, and all sample rankings contained at least one subjectivity factor. Furthermore, most sample rankings were based on features only, while only a quarter considered user experience. To determine whether user-experience-based popularity is comparable to online depicted popularity, 12226 user reviews were collected from two user review platforms by web scraping. The reviews were analyzed by natural language processing to obtain each scraped review’s sentiment. The aggregated sentiment data showed a slightly different popularity ranking; however, the leading tool remained the same. In addition to answering the research question, a more comprehensive popularity ranking was created. The ranking was based on multiple variables – the user sentiment, user count, and the online depicted popularity, whose inclusion was enabled by user review keyword analysis which proved that features are important to users and, therefore, feature-based rankings are worth considering. The more comprehensive popularity ranking also did not differ from the online depicted popularity. The results show an interesting contradiction – despite online rankings being exposed to subjectivity and having low reliability, the user data supports the same popularity ranking. The main benefits of this research are the raised awareness of the reliability of online sources, the conclusion that users do find features important, the observation that feature-based ranking matches the demand’s ranking, which points to a correlation between feature presence and the demand for the product, and the confirmation of the claim that popularity does influence consumer decisions as the most popular builder has the most users. The research limitations are the significant difference in the number of reviews each builder has, Capterra’s review platform’s review construction which induces the tendency of the reviews’ sentiment to be mixed, and the incompleteness of the feature detection keyword Annija Vītoliņa 3 lists. The future work of this study can be related to the increase of the online ranking review sample and the number of builders considered throughout the