Reputable firms try to correct for the biases you mention [1]. We have more evidence for the claim that Putin is domestically supported than otherwise. Which makes sense, given he's stabilized and relatively-competently managed their economy [2]. (Cf: Oil prices are up.)
[1] http://www.pewglobal.org/2017/06/20/president-putin-russian-...
[2] http://www.atlanticcouncil.org/blogs/natosource/what-is-the-...
The problem with that is you hve to measure how those biases change; even if you somehow of a good baseline adjustments, when you get new better numbers is it the regime being more popular or people being more afraid that criticism of the regime will be detected and punished?
Also, the survey you link doesn't actually describe any method of correcting for biases; implicitly, I suppose you could say that the art of options (some vs. a lot of confidence) and how they are treated (only “a lot” really treated as a positive signal) is a weak effort to mitigate the of bias at issue.
I agree that the data are imprecise. It is impossible to confidently answer "is Putin more popular today than a year ago?" But they are good enough to answer "does at least half the population support him?" Contrast that to oppressive regimes in Egypt or Venezuela, where the regimes do not appear to have popular support.
Independent polling showing strong support for the existing leadership in a country that has a long history with (perhaps) brief interruptions of being a heavily monitored authoritarian state with both overt and covert reprisals, not infrequently including fatal ones, against dissidents cannot be trusted for reasons which ought to be obvious.
Which is not to say Putin doesn't have strong support, just that polling can't be relied on to assess that.