It is an indispensable job to objectively evaluate academic institutions in daily scientific administration and webometrics-based evaluation technique has been increasingly becoming an effective way to fulfill this task besides traditional bibliometric indicators. In order to investigate the impact of different analytical strategies on the web influence of institutions, 342 universities in Mainland China were selected to compose the sample set to compare the reliability of different evaluating indicators on the web influence of institutions generated from different strategies of web data collection. Finally, the statistics shows that the numbers of educational domain names (RD_EDU) that linking to university websites, as well as the numbers of inlink on department level (Linknet) that university websites receive are significantly correlated (the average coefficient is nearly 0.8) with several transitional university ranking lists. Moreover, the correlation strengths are higher than that between prevailing web-based ranking lists (Webometrics Ranking of World Universities and uniRank) and traditional ones (ARWU, CARK and CUAA).