It’s been quiet. Too quiet.
This past week I attended four sessions with other SAP Champions [see which] and Jerry and Katarina Nonhebel Kati, looking at assumed unanswered questions, steering the results of Google search algorithms to provide better results for seekers, whether that be closing an old question, flagging one or more answers as correct/good, and deprecating others as unsolvable.
That was a mouthful. Rather than adopt my frequent persona of ABAP Detective, this will be a straightforward event reporting travelogue (in the video room road-trip adventure level only). In a brainstorming session a while back, I threw out the idea of an “answer-thon”, or virtual meet-up, to resolve a particular community challenge, as a more fun approach than individually looking at spreadsheets full of links and one-way comments (yuk). Once I opened my yap, I owned the idea and with execution help from SAP we set up 4 Zoom calls.
We had tried doodle polls to get people to “sign up” and as everyone has lots of things going on, it almost didn’t start. Alternative calendar invites worked better, with 3 to a dozen folks joining on one call or another. Jerry Janda led the display and conversation by scanning the given lists of questions, looking at the SAP Q&A site, then getting a consensus on actions such as closing, or maybe moving a comment to an answer.
Anyone who has done moderation, or participated in community discussion boards knows healthy useful content needs pruning and/or fertilizing to avoid stagnation or worse. Budget/marketing decisions aside, SAP community members have multiple ways to contribute content, from answering/commenting, to blogging, to writing up wiki pages. And that’s part of the dilemma this little event was going to run into, where good answers might already be found. Then there’s old content that despite the algorithms’ designers intent, lead people to an old answer instead of a new way of doing software/business as provided by evolved technology (and code).
Let’s start with a random example that was left unsolved by our team. The found question concerned “year end close”. So, a clearly important business question, yet one that has a lot of supplied rules and techniques already (country tax laws, best accounting practices, 101-level business processes). I shrank the first page I got back from Google, for viewing a lot in this small blog format. Will others get differing results? Probably, but I’ll go through mine anyway, then hopefully show why this matters. The first couple hits are SAP Help, officially, the second showing a 25 page PDF available for your reading pleasure (yours, not mine).
Vanilla search results for “year-end closing site:sap.com”
Layered below these include 2 sap.com site pages–(1) a blog post and (2) and question (with 2 answers). They are from 2008 and 2009, respectively. While a finance expert could review that post and determine whether the content is still relative, the question/answer chain cannot be so easily resolved, in my view.
Which leads to the question, is this level of effort really meaningful, or does someone need to gin up a state machine and auto-magically close out all questions on obsolete versions or menus, and save us much time.
The plan was…
- look at the title, the tag, and the date
- read the q and any answers
- note any ratings/feedback on the answers
At one point, I documented this decision tree, or triage, as a starting point for acting on each question. It should be a flow-chart/process map, hint-hint.
- does the subject make sense with the question?
- can the question be answered with the available info?
- are any answers close enough to be right?
- is there a similar question with a good answer?
- do the related questions apply/excel
- is there an FAQ in the wiki, or should there be?
- does the answer require more depth such as a blog or wiki?
- does the topic cover a deprecated or obsolete area?
- could/should the question be referred to an expert?
How it worked
Overall, the initial vision of a hack-a-thon like event, where you close bug reports or otherwise clean up a backlog of some data dam, was surprisingly close to actuality. The last day was bad for me, with 2 PC glitches taking me offline for longer than I’d have liked. I won’t name names, as there wasn’t really a recording of the Champion’s sessions, but will than those who attended. If they choose to comment here, great, we can bring ideas to the community at large, if not, we can keep working behind the scenes.
Zoom was a pretty good compromise as a technical platform, based on familiarity; it would be nice to try this with MS Teams. I think a Google meet is out of the question, but who knows. At least it won’t be Flash-based.
We tackled nearly 40 questions in 4 one-hour sessions, excluding individual time we spent updating the sheet or contacting peers for insight. I got to ping Dipankar Saha not to mention Susan Keohan and the list goes on.
One major discrepancy we faced was the low reported “views” on the question pages. Some showed as few as 25 views, but why would Google suggest such as low activity level content as better than others? Needs to be researched, though I must assume the counts SAP displays has been reset, like a crooked odometer, within recent times.
This is not an actual screenshot of Jerry trying to get the browser tabs to line up:
The 2 main complaints about the challenge we took on was the age of some questions, as well as the breadth of topics. For me, the latter was not as daunting, not that I see myself as a know-it-all, but it’s an opportunity to learn more, or at the least, help another future sojourner.
Aged questions can be filtered, if that’s how people feel best equipped to contribute, especially with many posts having anonymized/locked-out accounts. Guestified, if you will.
As to tags and the question universe, my initial take on reviewing the sheet was there mostly unique tags out of the 100 URLs. A revised query of the analytics dataset is in order I presume.
Would community members beyond the small number of Champions want to meet up virtually to hack through topics like this? I quite enjoyed being the Ed McMahon to Jerry’s Johnny Carson; scaling this out to more people in more time zones is going to need some kind of growth planning. What do you think?
— Jim Spath (@jspath55) November 5, 2020
Digging into SAP Q&A history, I was rightly called out for teasing about someone’s grammar/vocabulary. Thanks. Uwe Fetzer !
And your are right, too 👉 "Never make fun of someone who speaks broken English. It means they know another language."
— 🇪🇺 Uwe, still @ 🏠 (@se38) October 23, 2020
Insert Koan here, about patience with newbies.