For the first article which started off with a low number of downloads, but with the most impressing increase, 900%, there was another 2 dowloads during the second week. A small number but still a 50% increase from the original number of downloads for the first 8 weeks. For the second article, which had a bigger number of downloads already after the first 8 weeks, there was a more "modest" increase in downloads during the first week of sharing, 258%, there hasn't been any new downloads at all the second week.
My reflection is that the effect of sharing research on social media is quite limited in time. Already after a week it has "disappeared" in the flow of information if no actions is taken. My recommendation is that if you as a researcher, or as a librarian, share OA articles on social media you can expect a real effect in number of downloads, but it will be short-term. Therfore there is an opportunity to duplicate or rephrase your post after a month or so. But do not overdo with the risk of beeing repetitive!