Business Accent

To BI or not to BI?

Sanjay Shah on when to initiate the BI journey

Sanjay Shah

A few years back when the ERP fever started, big cheques were written, and the top managements thought that all their worries are now over once and for all. ERP’s did deliver process integration, best business practices, common code etc ..., but the decision- makers did not get that they wanted ... they did not get Management Information Systems (MIS) for Decision-Making or BI (Business Intelligence) as it is called now.

Sitting on mountains of data, BI technologies will now deliver what management wants. But the question is, when should you start your BI journey?

My answer is: If you have already implemented your ERP, the sooner the better. Even If you have not yet implemented ERP, or are in the process of doing the same, it is very important to start the BI initiative immediately, as explained below.

The timing of BI

BI reports give management the information to run their business, to take decisions. Would it not be logical if the management thinks aloud on what is required by them to run their business? Shouldn’t this be available to the implementing partners of the ERP so that they can ensure that the processes they setup will ultimately give the various dimensions required for BI? BI should not be an after thought, else you are in for a big surprise.

Think clearly on what information you need to run your business. Put your expectations down in an Excel worksheet along with some sample data. Ask one of your analysts to generate simple excel pivot tables on legacy/ERP data. Rerun this process till you are satisfied with the content of the report. Don’t forget to articulate, visualize and prototype your information requirements. Don’t assume that your ERP implementation partner will read your mind and automatically give you what you need. Nothing can be further from the truth. Creating BI is an engineering problem. It requires you to state clearly what you want and how you want it, and all the systems down the line have to be setup accordingly. And if you don’t do it, you are in a mess … you get data and not information !

Timing the introduction of BI in your organization is a very tough decision to take. It’s like getting married. A very handsome friend of mine was having trouble getting married … either he was busy studying, or he had just got a job, or he was not making enough money, or he was too busy with his career. However busy you are, you have to find time for what is important.

In my opinion beware of the following mistakes and potential delays in your organization:

• I shall do BI once my data is perfect
• I shall do BI once all my ERP modules are implemented
• I shall do BI once we are out of recession/when I have more funds
• I shall do BI once I am a little free
• I have the world’s best ERP, why do I need BI?
• I know everything, I don’t need BI
• I don’t need a third party BI, my ERP vendor gives me everything
• I shall explain all this in brief below.

(Here I am clarifying that my idea of BI is not just the big products like Hyperion, Cognos, Business Objects and such. But BI can be implemented very effectively even using simple tools like Excel Pivot Tables, simple and easy to use databases, etc. I am calling BI anything which enables a user to quickly generate (without much programming) a report which summarizes raw data and allows drill down, slice and dice and rotate, see data and compare data in a multi dimensional format and see various intelligent dimensions at just a click of a button.)

Mistake 1 : I shall do BI once my data is perfect

How does one ever know if the data is perfect? Suppose yours is a company with seven locations, five factories and you have an item master having about 50000 items. You have been tagging item groups to these basic items, for example all bearing items are being tagged as ‘Bearings’ etc. To know your stocks you generate a report location wise, another report factory wise. You generate different reports summarized at group level, and another showing itemized details and so on. I have actually seen 25 different types of inventory reports having been written and printed out month after month. The stock variance report (which measures changes of stocks at two points in time) was being manually generated (in excel) and was available only at a summary level. The users spent days trying to identify which items were causing the variances.

When I asked the key users why they did not introduce simple BI reports, I was told that the data was not yet ‘perfect’. I requested the IT manager to give me the detailed stock report (at an item level). I quickly converted this into a database table and made an Excel Pivot table of this. When I summarized this at a ‘group’ level, promptly excel showed me a group called ‘(blank)’. On clicking the detail button, I found that 55 items had not been grouped. In just one hour I was able to replicate the functionality of all the 25 ERP reports! With just a little more effort, I was able to give them a variance report comparing the total stock of the current month wrt to any previous month. The total variance was further analyzed into quantity and rate variance. We added various other dimensions to this report to give a very rich user experience.

• MOTS (Moral of the Story): You can use simple tools to not only get great information, but also use it as a cleansing tool. And remember, that data will never be absolutely perfect. Creating and correcting data is a cycle which never ends.

Mistake 2: I shall do BI once all my ERP modules are implemented

One of my client companies had a grand IT plan. A very popular ERP was chosen, a good implementation agency was chosen. In two years, the ERP would roll out throughout all the companies and plants of the group. It was decided to first complete the ERP implementation, and then work on BI. This would mean that BI would be done after two years or so. When I discussed this scenario to a friend of mine (who is a veteran in ERP implementation), he said that there is no such thing as ‘completing an ERP implementation’. ERP vendors will keep coming up with new ‘patches’, new functionalities and new versions so that you are perennially in a state of implementation or reimplementation.

Zoom into the Accounts Receivable department. The ERP went live there in about four months of the start date. Now that this department was on the ERP, the management was all over them demanding a huge amount of analysis and operating info. For practically two years, the departmental staff worked day in and day out to satisfy the demands of the management. The last time I counted, the line printers were printing about 15000 pages each month, being distributed physically to various branch offices and sales offices around the country. The reports were being printed customer wise, location wise, and responsibility wise. The reports were being printed in a summary format, a detailed format, and various such formats.

All this because BI was to be introduced only after the ERP was done. Now I fail to understand what the introduction of ERP in the Production Planning has anything to do with introduction of BI in the Accounts Receivable department.

Fortunately some sense prevailed. A single well designed BI report was able to meet all the above requirements. A simple email utility sent the BI report by email to all the branches and regional offices (containing only their slice of data). Pro-active reminder letters got sent to all the customers. The payback period of this small BI module was 1.5 (hold your breath) days! And it is green technology. We saved printing of 15000 pages every month!

• MOTS: Work on ERP and BI in a parallel mode and not in a serial mode for maximum effectiveness.

Mistake 3: I shall do BI once we are out of recession/when I have more funds

In a recession, as we are all in now, the manufacturing activity reduces and sales reduce. But it rarely reduces the work of the information worker. Recession does not reduce the number of items in the inventory or the number of general ledger codes! He still has to report on 50000 items of inventory, and 1000 GL Codes. There is a huge pressure on funds. People lose jobs, and those who remain suffer a burn out. They burn the midnight oil for meeting reporting and statutory deadlines. The management introduces a blanket ban on all new investments. And with this, also halt the proposed investments in BI.

I feel that a recession is the best time to improve and strengthen the internal processes. BI investments are not huge. If they are done at the right time, they will yield great returns, not only during the recession, but when the boom time comes you will be able to work much leaner. Remember that the cost of a BI initiative is not the cost you spend on the BI tools and technologies. Rather it is the opportunity cost which you pay because you did not implement BI.

Let me give you an example. A client of mine is in the auto component sector. The demand schedules given by their OEM customers keep changing. They did not have a mechanism by which these constantly changing schedules could be tracked in their popular ERP. As a result there was much more being produced and dispatched, leading to a lot of stock at the transporters warehouse located near the OEM’s factory. As the finished stock, once dispatched and invoiced, was reduced from the stocks, there was no track of this pipeline stocks. The Account Receivables were piling up as the customer refused to pay for the stocks they had not inwarded. Everyone was running in circles.

A simple tracking module with a lot of BI reporting on pipeline stocks, aging of such stocks, etc., was instrumental in completely eliminating the pipeline stocks. The client was able to recover the investment in days, and it freed a huge amount of working capital. The client now calls the BI investment as salt in his food!

There are various flavors of BI tools available in the marketplace. It is not necessary that when you implement BI, it has to be an enterprise mode BI with all the bells and whistles. A simple, incrementally build-able solution is good enough to start with, as long as it delivers what the management wishes. Once that is successful you will have a greater buy-in and more funds to go in for a larger solution.

• MOTS: BI can actually help you reduce you costs by unlocking your investments in inventory, account receivables and understanding your product and customer profitability.

(To be continued next week)

Sanjay Shah is the CEO of Prosys Infotech Private Limited, a Pune based company specializing in developing BI solutions on the Microsoft BI platform. He can be contacted at sanjay@prosysinfotech.com