We have a client who has almost 1,000 custom reports (mostly crystal) in E9 looking at an upgrade. We don’t think they really know which reports are truly important. The Cirrus tool only tells you when the report was last modified, not LAST RUN.
Is there any way to tell when all the defined reports were ever actually run by a user to say something like “Any report you haven’t used in the last 3 years should be parked and analyzed in the future only if someone needs it” versus spending tons of time and effort uplifting tons of reports that may never get used again.
Nothing native, could create a bpm to record in a UD table when a BAQ/Report was last executed. After running that for a little while do some analysis then clear out the UD table that was used.
I would take the opportunity to replace some of these reports with Dashboards or EDD in .300 and up. An efficient organization doesn’t run on reports as much.
You could also rename all of the questionable ones and see which ones are REALLY needed.
I was sitting through some of Brian Conner’s EDD sessions and with .400+, it’s getting a lot of upgrades! First of all, reporting of Epicor Data is free. You’ll be able to combine BAQ, MSAS, and REST data sources. The widgets can be embedded in Forms so it would filter the graph in context of the hosted screen. IoT data widgets displaying “real-time” data. You can drill down to the source data (BAQ I assume) to get to the original record. Finally, the current filtered record set can be downloaded to Excel. I am very impressed and it looks like it’s getting a lot of development attention too.
The problems I see with EDD, and correct me if you think I"m wrong:
Embedding the widget in the forms is a cool feature, but when used in practice (as we saw with Microsoft), can bring performance to its knees. Imagine an EDD view tied to Sales order entry that shows historical sales of an item as you enter it on a line. Imagine how much data and hit to the database would occur each time you move to a new line with a database that might have a million + Invoice history lines to pull the sales data via the baq?
2)The source data link is now giving it the ability that ‘regular’ dashboards always had
You could export in 10.2.300.
As far as I know there is no cross filtering, no ability to create something akin to publish and subscribe or a dashboard where one EDD view relates or interacts with another (you need EDA for that)
STORY TIME: Too many years ago (Almost 28 to be exact), I worked in an old corporate IT dept where we ran nightly reports… the Computer operators, removed the reports from the 900 Line per minute printers, decollated them (yes, decollate is a word), and stacked them by department, and every morning, the “mail room” staff came by and picked the reports up and delivered to whomever printed the reports. SOME of the reports were auto printed every night. SOME of those reports were never examined (imagine that).
About once every 6-8 weeks, we told the mail-room staff to NOT deliver the reports. instead, they delivered a note that instructed the person to come by and pick up their reports and any reports left behind would be removed from the nightly runs.
It was amazing the number of reports that we stopped printing nightly every time we did this.
Something to consider here though is the go-forward with kinetic you’ll be talking a lot of async calls so while that is off spinning it’s wheels you’re still working you’re not locked in a thread. Just a different way to think about data retrieval than what we deal with today.
Still, the only proper way to display real time analytics ‘in context’ and not put unnecessary strain is with some kind of Datawarehouse that pre-aggregates the data into cubes that require no horsepower or time to retrieve. SAP does this with their products because instead of SQL, they use HANA which means the entire database run’s in RAM and can ‘slice and dice’ these numbers as quickly as SQL Server Analysis Services does.
But using straight SQL queries (via BAQ’s or REST) is going to be highly inefficient and potentially crippling if it isn’t analyzed by a professional before being put into place.
It’s the same reason we see clients with customizations (that we didn’t do) that make SO Entry go from a 4 or 5 second open, to a 45 second open, or hitting a new line on a PO takes 30 seconds because of BPM’s firing that are retrieving tons of data per line
Ok, so much for ‘as far as i know’ considering these potentially come from different queries with different aggregation levels, would be interesting to see how cross filtering works, unless it only works with one element at a time (ie highlight a particular customer, and see the job efficiency of that customers jobs, but would only work if the BAQ on the job data is already at the customer level)
You can’t really compare Tier 1 and Tier 2 ERP applications as equals though. You’re talking different scale at that point. But I digress not germane to the posted topic, but just something to be aware of in conversation for folks that might not make that connection.
SAP Business One uses HANA as well and is about 1/2 the price of Epicor and about 1/2 the complexity. Downside is it’s only good for the simplest of manufacturers as it focuses on distribution and finance.
Same here. I wasn’t impressed with EDD until Insights this year. I think it was my number one take away and something I’ll be looking into more this year.
The funny thing is Epicor abandoned SSAS when they moved from EPM (for good reason) to EDA. So companies will need another way to create SSAS cubes which is no cakewalk for the inexperienced.
