Isolating and Analytic languages tend to have a more rigid word order because they lack the encoded information of inflectional/derivational morphemes. As such, these types of natural languages rely on the word order or additional content words to convey meaning. Natural languages that allow for "free word order," or more accurately less strict word orders encode grammatical info (Case/Tense/Gender/Number/etc.) into each word, often relying more on these morphemes and agreement to convey the intended meaning. Even if word order is ignored, the same meaning would be conveyed by the included morphemes. These morphemes are what allow for a language to easily change its preferred word order.
To keep the structure and word order of an analytic/isolating conlang, one can focus more on the way that constituents come together to form noun phrases, verb phrases, and clauses. That said, natural languages can also be flexible and allow multiple, acceptable constructions, but "certain conventions" would be the most common pattern used native speakers. In English, one could say "the red, big dog caught the ball" and be understood although convention would dictate "the big, red dog caught the ball" as more acceptable. However, "the ball the big, red dog caught" would change the meaning entirely or be considered ungrammatical.
Having personally studied Mandarin Chinese and White Hmong (both strong analytic languages) I found that word order was very strict when dealing with the positions of nouns, verbs, adjectives, articles, and prepositions, but they had different approaches in how clauses could come together.
Slioussar, N. (2011). Processing of a Free Word Order Language: The Role of Syntax and Context. Journal of Psycholinguistic Research, 40(4), 291–306. https://doi.org/10.1007/s10936-011-9171-5
Gell-Mann, M., & Ruhlen, M. (2011). The origin and evolution of word order. Proceedings of the National Academy of Sciences, 108(42), 17290–17295. https://doi.org/10.1073/pnas.1113716108
