For me as a native speaker: yes, longer German words have syllables with secondary stress, and it is important to put the correct stress (pimary and secondary) on the right syllable.
Test:
Au(1)-to-ge-schäft(2)
is correctly pronounced. Whereas
*Au(1)-to-ge(2)-schäft
is a technically possible way of pronouncing this word, but it is definitely wrong. It sounds as if the person speaking was from Russia or Hungary. (Both have their peculiar and separate problems with finding the correct syllables to stress in German.)
Similarly
Ein(1)-tei(2)-lung
is correct, whereas
*Ein(1)-tei-lung(2)
sounds completely idiotic, and I can even not imagine a native speaker of what language would be in danger of pronouncing it this way. If any, it would be something as far away as Mandarin.
You can do this test with every longer word: Rattenschwanz, Gebührenordnung, Feuerwehrleiter, Hundekacke, Kamelhaarjacke, Suppenschüssel, Oxymoron, Ausspracheregel, Poesiealbum, Onomatopoeie, Auspuffmuffe...
Later thoughts: is there a 'potential secondary stress'?
After some more in-depth discussion in the comments e.g. around words such as Stadtrundfahrt, as well as with respect Janka's separate answer, I would admit that it can justifiably be claimed that Stadtrundfahrt can also be seen as having one single syllable stress (*Stadt(1)-rund-fahrt), i.e. one can see the other syllables as equally unemphasised.
However, there seems to be at least a potential secondary stress bearing syllable; which becomes visible as soon as one tries to put a secondary stress on one of the remaining syllables. Put it on the wrong syllable, and you get a clear mistake: Stadt(1)-rund-fahrt(2), whereas put on the correct syllable, everything sounds okay: *Stadt(1)-rund(2)-fahrt.
Consequently I would instroduce here the term of a "crypto-secondary stress location", "hidden secondary stress location", or "potential secondary stress location".