Michael Eriksson's Blog

A Swede in Germany

Some more experiences with eCommerce and poor web-design

with 5 comments

Since I have spent (and intend to continue to spend) a lot less time traveling, I have just ordered a desktop computer to get more comfort over my laptop.

This provided several good illustrations of how poorly thought-through many web-shops are.

For instance,* during the actual order stage, I found that copying my VAT-identification into the corresponding field led to an unspecified error—allegedly, something was wrong, but no word was given as to what. A bit of experimentation revealed that because I had not typed the VAT, the field did not recognize that an entry was present… This is idiotic on at least three counts: Firstly, this is the type of information that should be copied as a matter of course, to reduce the risk of accidental errors (and work needed). Secondly, a good developer would not have let himself be fooled by something like that.** Thirdly, a reasonable error message should have been given, e.g. that a mandatory field was empty. This would have made the error search much faster and would have avoided red herrings like a syntax error or an accidental copying of the wrong value (or an incomplete copying of the correct value). The error message displayed also showed the design error of appearing (only) at the top of the page, instead of next to the field. This was especially bad because the top of the page was not visible without scrolling when the field was visible… (However, at least the field was actually marked red, so that the user knew to search for an error message—I have seen even this be left out on some other sites.)

*At http://www.cyberport.de

**Most likely, a heuristic was used that “if the user has typed something, the field is not empty”. This is highly naive and the (easy) check whether the field actually was empty would be much better.

Similarly, there was one of those idiotic* “please re-type your email address” fields. Of course, I just grabbed the original entry and pasted it—and nothing happened. Apparently, instead of realizing that this type of field is an idiocy, the designer had decided to block copy actions to force a re-type. To boot, this was done without any discernible error or warning message.

*Email addresses, too, are best copied from a fix source—not typed. If it is copied, there is no risk of a mistype and the “re-type” field is a pointless time-waster. Most non-copiers will likely rely on auto-complete, which will almost always either give the correct result or the wrong result twice. Again the “re-type” field is a pointless time-waster. For those who do type, the clear majority can be expected to either type and double-check sufficiently carefully that the address is correct in the first field, while those who do commit an error will usually do so due to a memory error, which will usually be repeated in the “re-type” field… Again, this field is a pointless time-waster. (A better approach could be to e.g. put a warning text next to the first field, to indicate the communication problems that could ensue if the address is mistyped and “please double-check it”.

Earlier, I had visited a number of category pages from one of Germany’s most popular physical electronic stores.* This with an eye on looking for other things that might be interesting, the store(s) having a very wide selection of products, be it on- or off-line. Not only did I have to jump through hoops to get to these pages,** but once there, they were all empty… Whether this was due to an internal error or an unprofessional reliance on e.g. Google***, I do not know. What I do know is that I wasted a fair amount of time, bought nothing, and definitely will not return in a hurry.

*Conrad resp. http://www.conrad.de

**There were usually several clicks and a lot of scrolling needed (instead of the one click that should have been needed), because the original links did not lead to the category pages—but to information pages that contained a link to the real category page somewhere towards the bottom.

***Google (and a few other companies) provide extensive APIs that can facilitate web-development. For an online store, it should be a given that these are installed locally. However, some developers fail to do so, and instead rely on versions running on Google’s (or whoever’s) servers. This brings problems both with reliability and user privacy, and I have blocked some of these servers to protect myself from privacy violations.

The search criteria in several stores were abysmal*, missing even basics like the ability to filter computers based on e.g. OS (specifically, no** OS), amount of RAM, and similar. Typical sets of criteria were brand (rarely interesting***) and price (interesting, but not enough) and possible something else of lesser import (e.g. whether shipping could take place now or only in two days time). This resulted in result lists of dozens to hundreds of entries that had to be manually filtered. (With the effect that I looked through the first one or two pages, foregoing the many entries on later pages entirely…)

*Including http://www.cyberport.de

**As a Linux user, I do not want to pay extra for a Windows installation that I am just going to remove later. Of course, even among those content with a pre-installed OS, the question of which OS is often quite important. As an aside, the proportion of computers that still come with a pre-installed Windows is depressing—the year is 2019, not 1999, and it should be a given that a Windows installation is optional.

***While some might have a brand preference, it is usually far more important what characteristics the computer has, and in those rare cases where someone is justified in looking at the brand first, he would be better of going directly to the manufacturer website (for research, if not for the actual purchase; of course, after research, he could just search by product number and would never need the brand). A possible exception is a means to exclude some brand; however, this was never present. (Except by selecting all brands, and then de-selecting the one—with possibly disastrous time waste as the result. Cf. immediately below.)

A particular annoyance was the slowness that came through attempts to be interactive—confirming my observation that the more interactive and “helpful” a website tries to be, the slower and less helpful it tends to become. Notably, changing any filter setting leads to an automatic re-load, which implies a re-search or re-filtering server-side, which implies a considerable delay until the page is available again. However, it is rarely the case that the user only wants to change one filter setting,* and it would usually** be better to have him toggle the reload manually after making all changes. Consider e.g. prices on http://www.cyberport.de: Per default, they ranged from a few tens of Euros*** to many thousands. Naturally, I wanted to trim both values to, respectively, ensure that I got something actually usable and did not pay a fortune for something I did not need. However, to trim the range to e.g. 200–500 Euro, I had to change two filter settings. Both caused a reload with a significant loss of time.****

*Unless, obviously, the number of settings is too limited to begin with…

**One exception is when the one choice alters what other choices are available. This was not the case on these websites, however. (And when it is, it is usually better to pre-load such alterations in a manner that allows a client-side change of filter options without reloading the actual results from the server.)

***Presumably, either non-computers misleadingly put in the computer category or extreme mini-computers (Raspberry PIs or similar).

****To boot, the settings were not even input fields, but some type of weird bar, where the user had to move the ends of the bar until approximately the intended values appeared.

Excursion on email:
While a bit off topic, I note that Cyberport provided yet another example of the grossly unethical practice of not having a means to provide an email address without also consenting to spam, insteading forcing the user to revoke consent at a later time. (Of course, not providing an email address at all is not an option.)

Advertisements

Written by michaeleriksson

March 26, 2019 at 9:37 pm

5 Responses

Subscribe to comments with RSS.

  1. […] a text from earlier today, I referenced several web-sites. I deliberately did so without linking and mentioning just the […]

  2. […] to and after writing the previous installment ([1]), I tried to use the German Bauhaus for some major purchases for and/or some works on my apartment, […]

  3. […] of my recent (and past) problems with eCommerce (cf. e.g. [1], [2], [3]) were worsened by a fundamental flaw in the typical* legal view of how a (B2C) purchase […]

  4. […] a further follow-up to recent writings ([1], [2], [3]) I have to categorically advice to stay away from Cyberport. What has transpired here is […]

  5. […] deliveries already discussed in [1] and [2]; around payment methods, web-design, etc. in e.g. [3] and [4]; and some other aspects in an up-coming […]


Leave a Reply to The perversity of offers in B2C commerce | Michael Eriksson's Blog Cancel reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s