Business Intelligence
Discussion
Right, we have two lots of data - STOCK and PRODUCTS (with sales)
I want to cross reference them with each other so I can query the STOCK and find out which 'products' creates the sales
I also want to Query the products to find out where sales might be taking place
Many products are use multiple lines of stock
With 3500 items of stock and 150,000 'products' things are getting a bit complicated to keep an eye on.
Anyone here with this kind of experience? We really want to be able to drill down through our data in order that we can monitor sales (or lack of) of the 3500 items of stock. We want to be notified of changes in sales and all sorts of things we don't yet know we need to know!
I've heard of IBM's COGNOS but at £10k a pop is a little on the pricey side! I think I could do it with (and probably will to test the waters) 2 sets of data - Products and stock and VLOOKUP against each other.
I want to cross reference them with each other so I can query the STOCK and find out which 'products' creates the sales
I also want to Query the products to find out where sales might be taking place
Many products are use multiple lines of stock
With 3500 items of stock and 150,000 'products' things are getting a bit complicated to keep an eye on.
Anyone here with this kind of experience? We really want to be able to drill down through our data in order that we can monitor sales (or lack of) of the 3500 items of stock. We want to be notified of changes in sales and all sorts of things we don't yet know we need to know!
I've heard of IBM's COGNOS but at £10k a pop is a little on the pricey side! I think I could do it with (and probably will to test the waters) 2 sets of data - Products and stock and VLOOKUP against each other.
One post does not make a spec 
If you think VLOOKUP (or whatever) is sufficient for your needs then there's no need to over-complicate with proprietary BI 'solutions'. With a firmer idea of what you want to do, you could try somewhere like rentacoder and get some external assistance. Were you to get more sophisticated, there are plenty of opensource (ie relatively cheap) BI products on the market (Pentaho is a good one) if you think Cognos would be too pricey. I work exclusively with a more expensive (and hence much more superior *cough*) product but it all comes down to what do you really want to do?
Oh, and miserable get that I am, the dificult bit with building any BI project is getting the data in, not getting the information out.

If you think VLOOKUP (or whatever) is sufficient for your needs then there's no need to over-complicate with proprietary BI 'solutions'. With a firmer idea of what you want to do, you could try somewhere like rentacoder and get some external assistance. Were you to get more sophisticated, there are plenty of opensource (ie relatively cheap) BI products on the market (Pentaho is a good one) if you think Cognos would be too pricey. I work exclusively with a more expensive (and hence much more superior *cough*) product but it all comes down to what do you really want to do?
Oh, and miserable get that I am, the dificult bit with building any BI project is getting the data in, not getting the information out.
What we want... I'll start with what we have - a list of 3500 items of stock
We also have lots of sales data - more every day that includes what stock is sold, the product that created that sale, the source of the sale, price of sale and so on.
We really need to know:
Which STOCK simply rarely sells so we can either bin them or create new PRODUCTS to incease sales
Which PRODUCTS don't sell so we can concentrate on the ones that do.
Where which products are being sold and so forth so we can really gain an insight in to the what's and why's.
we particularly want to be able to monitor groups of products for sales changes and then be able to drill down into the info if / when we see a change. For example (for me, not you) PAINTS > household > colours. So if general sales of paints are down we can see overall it's household paints as opposed to commercial and which colour has collpsed sales. Of course we want to know about increases in sales too!
There's probably a million other things we would like to know as soon as we know what they are as we don't yet know :S
The main thing that puts me off cognos is that the £10K is for 25 users - we have only 4. I really need a friendly expert in this field as we're flying in the dark without a night rating and running low on fuel
We also have lots of sales data - more every day that includes what stock is sold, the product that created that sale, the source of the sale, price of sale and so on.
We really need to know:
Which STOCK simply rarely sells so we can either bin them or create new PRODUCTS to incease sales
Which PRODUCTS don't sell so we can concentrate on the ones that do.
Where which products are being sold and so forth so we can really gain an insight in to the what's and why's.
we particularly want to be able to monitor groups of products for sales changes and then be able to drill down into the info if / when we see a change. For example (for me, not you) PAINTS > household > colours. So if general sales of paints are down we can see overall it's household paints as opposed to commercial and which colour has collpsed sales. Of course we want to know about increases in sales too!
There's probably a million other things we would like to know as soon as we know what they are as we don't yet know :S
The main thing that puts me off cognos is that the £10K is for 25 users - we have only 4. I really need a friendly expert in this field as we're flying in the dark without a night rating and running low on fuel

If Excel is too basic and Cognos too expensive, you should be able to constuct a star schema in any database and from this build a cube using SQL Server Analysis services. Costs would be a standard edition SQL server license.
The complexity of this will depend on how and where your source data is stored - as someone has said the difficulty is always getting data in.
You could then get Excel to connect to your cube and drill down and see which products and stock items have sold well at different points in time (this is all built into Excel 2007). If you do things like group products into categories you can drill down on this as well. You should also be able to set something up to notify you when things happen, e.g. if the stock levels of stock items that make up a major product are getting low. Feel free to PM me if you want to talk through this in more detail.
The complexity of this will depend on how and where your source data is stored - as someone has said the difficulty is always getting data in.
You could then get Excel to connect to your cube and drill down and see which products and stock items have sold well at different points in time (this is all built into Excel 2007). If you do things like group products into categories you can drill down on this as well. You should also be able to set something up to notify you when things happen, e.g. if the stock levels of stock items that make up a major product are getting low. Feel free to PM me if you want to talk through this in more detail.
WH0Sace said:
Right, we have two lots of data - STOCK and PRODUCTS (with sales)
I want to cross reference them with each other so I can query the STOCK and find out which 'products' creates the sales
I also want to Query the products to find out where sales might be taking place
Many products are use multiple lines of stock
With 3500 items of stock and 150,000 'products' things are getting a bit complicated to keep an eye on.
Anyone here with this kind of experience? We really want to be able to drill down through our data in order that we can monitor sales (or lack of) of the 3500 items of stock. We want to be notified of changes in sales and all sorts of things we don't yet know we need to know!
I've heard of IBM's COGNOS but at £10k a pop is a little on the pricey side! I think I could do it with (and probably will to test the waters) 2 sets of data - Products and stock and VLOOKUP against each other.
It should be straightforward to either import or link to MS Access.I want to cross reference them with each other so I can query the STOCK and find out which 'products' creates the sales
I also want to Query the products to find out where sales might be taking place
Many products are use multiple lines of stock
With 3500 items of stock and 150,000 'products' things are getting a bit complicated to keep an eye on.
Anyone here with this kind of experience? We really want to be able to drill down through our data in order that we can monitor sales (or lack of) of the 3500 items of stock. We want to be notified of changes in sales and all sorts of things we don't yet know we need to know!
I've heard of IBM's COGNOS but at £10k a pop is a little on the pricey side! I think I could do it with (and probably will to test the waters) 2 sets of data - Products and stock and VLOOKUP against each other.
From there you could either learn to create queries and reports graphically or a template system could be set-up so that end users can get the information directly. Reports are a straightforward job to create so they could be created as users request information. I can do such a system fairly easily. PM me if you want, I can send screengrabs of what is possible (I did a pc building system for a local company a few years ago which was rather involved, rigid stock tracking of individual batches of stock (ie per delivery))

purplepolarbear said:
If Excel is too basic and Cognos too expensive, you should be able to constuct a star schema in any database and from this build a cube using SQL Server Analysis services. Costs would be a standard edition SQL server license.
The complexity of this will depend on how and where your source data is stored - as someone has said the difficulty is always getting data in.
You could then get Excel to connect to your cube and drill down and see which products and stock items have sold well at different points in time (this is all built into Excel 2007). If you do things like group products into categories you can drill down on this as well. You should also be able to set something up to notify you when things happen, e.g. if the stock levels of stock items that make up a major product are getting low. Feel free to PM me if you want to talk through this in more detail.
+1The complexity of this will depend on how and where your source data is stored - as someone has said the difficulty is always getting data in.
You could then get Excel to connect to your cube and drill down and see which products and stock items have sold well at different points in time (this is all built into Excel 2007). If you do things like group products into categories you can drill down on this as well. You should also be able to set something up to notify you when things happen, e.g. if the stock levels of stock items that make up a major product are getting low. Feel free to PM me if you want to talk through this in more detail.
i had hoped the OP already had SQL Server
you can download a trial here http://www.microsoft.com/sqlserver/2008/en/us/try-... but of course you will need some hardare to install it on..
WH0Sace said:
I'm happy to purchase anyhardware requried.
you'll need to buy a Windows machine, as the only Mac db i can think of is Filemakeri would suggest ffenics, at only £295 per full licence (£100 per run-time) it's not going to break the bank
and you can download an evaluation version to see if it works aswell
if your data is in excel, then you could import it or connect via ODBC ?
Edited by Man-At-Arms on Wednesday 3rd March 12:57
Man-At-Arms said:
WH0Sace said:
I'm happy to purchase anyhardware requried.
you'll need to buy a Windows machine, as the only Mac db i can think of is Filemakeri would suggest ffenics, at only £295 per full licence (£100 per run-time) it's not going to break the bank
and you can download an evaluation version to see if it works aswell
Oracle
MySQL
Postgresql
Edited by cs02rm0 on Wednesday 3rd March 13:04
purplepolarbear said:
If Excel is too basic and Cognos too expensive, you should be able to constuct a star schema in any database and from this build a cube using SQL Server Analysis services. Costs would be a standard edition SQL server license.
The complexity of this will depend on how and where your source data is stored - as someone has said the difficulty is always getting data in.
You could then get Excel to connect to your cube and drill down and see which products and stock items have sold well at different points in time (this is all built into Excel 2007). If you do things like group products into categories you can drill down on this as well. You should also be able to set something up to notify you when things happen, e.g. if the stock levels of stock items that make up a major product are getting low. Feel free to PM me if you want to talk through this in more detail.
+1 on this approach..... The complexity of this will depend on how and where your source data is stored - as someone has said the difficulty is always getting data in.
You could then get Excel to connect to your cube and drill down and see which products and stock items have sold well at different points in time (this is all built into Excel 2007). If you do things like group products into categories you can drill down on this as well. You should also be able to set something up to notify you when things happen, e.g. if the stock levels of stock items that make up a major product are getting low. Feel free to PM me if you want to talk through this in more detail.
The big question... If you dont have a database already for this "data" then how are you maintaining it? If its excel spreadsheets etc, then it maybe that you need some kind of sale order processing system in the first place to make sense of the various activities you are doing; this system will likely also come with pre-prepared reports out of the box thereby meaning you dont need to spend/waste time on writing reports that millions of other businesses take for granted.
I can remember when Business Intelligence was called EIS (Executive Information Systems)
There are a couple of Open Source suppliers out there. Pentaho and Jaspersoft.
When I worked for Gentia in the mid 90s I could have knocked up a system in about 30 minutes. It would have worked on Windows, Mac and Unix. Such a shame that not all bleeding edge development companies survive.
There are a couple of Open Source suppliers out there. Pentaho and Jaspersoft.
When I worked for Gentia in the mid 90s I could have knocked up a system in about 30 minutes. It would have worked on Windows, Mac and Unix. Such a shame that not all bleeding edge development companies survive.
WH0Sace said:
Right, we have two lots of data - STOCK and PRODUCTS (with sales)
I want to cross reference them with each other so I can query the STOCK and find out which 'products' creates the sales
I also want to Query the products to find out where sales might be taking place
Many products are use multiple lines of stock
With 3500 items of stock and 150,000 'products' things are getting a bit complicated to keep an eye on.
Anyone here with this kind of experience? We really want to be able to drill down through our data in order that we can monitor sales (or lack of) of the 3500 items of stock. We want to be notified of changes in sales and all sorts of things we don't yet know we need to know!
I've heard of IBM's COGNOS but at £10k a pop is a little on the pricey side! I think I could do it with (and probably will to test the waters) 2 sets of data - Products and stock and VLOOKUP against each other.
You mention Stock, Products and Sales but you don't mention customers. Do you know who you are selling to? If so that would make your data much more valuable by being able to predict who will buy what product next. Not trying to push our solution (it's around £500,000 and a sophisticated predictive modeling solution, so not within scope for you) but you might get alot more from your data if you know who your customers are and what they have bought in the past using some history (such as seasonal trends, product combinations etc) I want to cross reference them with each other so I can query the STOCK and find out which 'products' creates the sales
I also want to Query the products to find out where sales might be taking place
Many products are use multiple lines of stock
With 3500 items of stock and 150,000 'products' things are getting a bit complicated to keep an eye on.
Anyone here with this kind of experience? We really want to be able to drill down through our data in order that we can monitor sales (or lack of) of the 3500 items of stock. We want to be notified of changes in sales and all sorts of things we don't yet know we need to know!
I've heard of IBM's COGNOS but at £10k a pop is a little on the pricey side! I think I could do it with (and probably will to test the waters) 2 sets of data - Products and stock and VLOOKUP against each other.

Gassing Station | Computers, Gadgets & Stuff | Top of Page | What's New | My Stuff




