BI Deployment: Horizontal or Vertical?

Should a company look at a horizontal design for business intelligence applications or does going vertical make more sense? Read on By Sanjay Shah

The company where Mr Basu works, ABC Corporation, is a conglomerate of several different types of businesses operating under one single legal entity. Mr. Basu was the head of SBU-A till the past year. But now he heads the BI Initiative. Growth in the past couple of years has been good. The group has invested in an ERP system and various modules of the system are under different stages of implementation. The MD firmly believes the continued growth and profitability will depend on taking the right decisions at the right time and based on data rather than on gut feelings. For this he feels that the time has come to invest in BI Technologies and applications. He has engaged Mr. Waghmare, who is a BI Strategist for guidance, and internally the initiative is driven by himself and Mr. Basu.

Mr. Basu feels that the BI applications should be designed horizontally. But the MD feels that they should be designed vertically.

The financial year has ended well and there has been good top line growth, but the bottom line is under pressure. So the mood of the MD is mixed.

Lata (MD’s Secretary): Good evening Mr. Basu and a happy new year. the MD and Mr. Waghmare are waiting for you.(She winks). But be careful, the MD is not in a very good mood.

Basu: Thanks, Lata.

Basu enters the MD’s room.

Basu: Good evening, Sir. Good evening, Mr. Waghmare.

MD: Good evening, Basu. (Sounding a little irritated) We have exchanged so many emails on the basic BI strategy, i.e. whether to design applications horizontally or vertically. I see that you are still not convinced!

MD’s secretary rolls in a tray with three glasses of chilled kairi panhe (a delicious Maharashtrian drink made from raw mangoes).

Basu: Sir, I have talked to many CIOs in different large companies, and all are saying that we should design horizontally and not vertically.

Waghmare: Hey, I don’t understand what you are saying. What is horizontal and vertical in BI application design?

MD: Let me explain. The terms horizontal and vertical are our nomenclatures. By ‘horizontal’, what Basu means is that first we should build dashboards that give a big picture of all the major areas of business. In each area we do not go down too much into details. By ‘vertical’, what I mean is that you take up one critical area of business and build dashboards as well as analytical tools just for that area. With this approach, the senior users are able to see a big picture of the area using a dashboard, and the analysts are also able to drill down to the lower levels of detail if required.

Basu: Mr. Waghmare, I feel that it is very important to show to the top management, the overview of business on a daily level. With this they will be able to know whether our ship is going in the right direction. Later on, we can always make detailed reports. That way you keep the top team happy and they know that their BI spend is showing some results.

Waghmare: Mr. Basu, I do see the point which you are trying to make.

MD: But what I feel is that if we do not give analytical reports to the analysts, they will not be able to act upon the observations of the senior team.

Waghmare: What you are saying is also right. It is in fact a perennial debate. And I believe, it can be resolved only by trying out both the approaches observing the results.

Waghmare: (continuing): ABC Corp is a large organisation. Let’s start in SBU-A the horizontal methodology, and lets start in SBU-B the vertical methodology. And lets meet up in about 3 months’ time, and review the results after the first quarter ending.

MD: Yes, that’s a good idea.

The team briefly discusses the business outlook of the current quarter and after about another half hour they disperse.

After the first quarter ending, a BI implementation meeting is once again called by the MD. Mr. Basu comes for the meeting. He is wearing a flashy red tie and his face shows considerable confidence. He has been able to make an impressive looking dashboard (showing almost 20 different areas), which has been highly appreciated by the Board Members. He has also, discreetly, found out what the MD has been doing in SBU-B and has been told that they have been barely able to do some two areas. So, Basu has come prepared to take the kudos and he feels that he would now be
able to head the BI implementation across the conglomerate.

Lata: The MD is in a really good mood today.

Basu (surprised): Thanks, Lata.

Basu knocks and enters. The MD and

Mr. Waghmare are having a cup of tea.

MD: Tea, Mr. Basu?

Basu (accepting a cup of tea): We had an excellent presentation of our BI initiative to the Board members yesterday. I am glad to say that it was very much appreciated.

MD: Yes, that’s true. The board members did like the look and feel of the dashboard, the colour combinations and the ability to see the big picture and drill down to the SBU level and fix responsibilities. Mr. Basu, I must say that I am greatly impressed.

Basu: Thank you, sir. So now you must be convinced that my approach is correct.

MD: I did not say that!

Waghmare: MD, let’s see your dashboard.
The MD opens up a dashboard and it shows up on the projector screen. It is a simple dashboard. Two major elements of Working Capital viz Account Receivables and Inventory are shown on a screen. The user is able to drill down to the SBU level details of the Working Capital. They are able to see a simple KPI of Working Capital as a Ratio of Sales. The previous year numbers are also available for comparison. There are some additional sheets showing Item level details of Inventory, and customer level details of DSO.

Waghmare: What I can see is that in SBU-A, where the horizontal method was implemented, the working capital ratio to sales has more or less remained in line with the previous year. However, in SBU- B, where the vertical method was adopted, the working capital ratio to sales has reduced by almost 23%. This had the effect of releasing almost Rs100 crore locked in working capital.

MD: That’s right.

Waghmare: Mr. Basu, can you open your P&L dashboard and tell me the ratio of interest to sales ?

Basu: The ratio of interest to sales for SBU-A has been about 6% in both the years, while in case of SBU-B it has dropped from 6% to about 4%.

Waghmare: The profitability has therefore increased by about 2%, even though the gross profit ratio has remained the same. Well that’s quite impressive.

Basu: But you see that is not very important. Today in SBU-A, the numbers of orders, dispatches, inventories, sales, cash on hand, AR, collections, etc, are all available daily. I have created an automated subscription, which sends all the key numbers to all the members of the board at 9 am each day.

Waghmare: That is true, but with all the numbers available, yet the profitability has not improved.

Basu: Well the circumstances have been challenging. There is recession all around, people are not paying on time etc.

Waghmare: But SBU-B has also been operating in more or less the same environment. How did they manage to release over 100 crores of working capital and improve their operating margin by 2% ?

Basu (sweating a bit in spite of the chilled AC temperature): I..I don’t know. I am sure that the MD must have put a lot of pressure on SBU-B.

MD (laughs): As Jan Carlzon once said, “Without information, people cannot take responsibility; with information they cannot avoid taking it.”

MD: Basu, just as you discreetly tried to find out what we were up to in SBU-B, I also tried to find out what your team was up to in SBU-A. I found out that while SBU-A also has all the information at the top level, the actual executing team did not have the required information to act on. Due to this, they were not able to achieve the results which my team achieved.

Waghmare has immediately understood the MD’s strategy, but is trying to make Basu understand the same by asking probing questions.

Waghmare: I am not able to understand. Both the SBUs are using the same ERP. So how is it that SBU-A did not have access to information?

MD: A very valid question, Mr. Waghmare. Let me explain. Instead of focusing the time of my BI team on the different areas of dashboard, I first asked myself which is the most painful area. I and my team came to a conclusion that Working Capital management was the most painful area and a lot of our investment is blocked in the same. Even a small reduction in working capital would release a lot of money for other critical requirements, and would also reduce our interest cost.

Waghmare: Very true.

MD: So I first created a dashboard of two important Working Capital components viz the Inventory and the Account Receivables. All the numbers, which I needed were available from the Balance Sheet and from the P&L Account.

Basu (a bit cheesed off): Yes, that’s the approach we also took. We have also created most of our dashboards from the Balance Sheet and the P&L Account.

MD: Once I saw the trend, I realised that Inventory and AR ratios to sales were both creeping up. So I called in my ERP expert and tried to explain to him that I needed the Inventory and AR data at a lower level of detail. I needed the Inventory Turns at the item level, and I needed DSO at the customer level. We tried to calculate the Inventory Turns for a few items from the ERP screens, and similarly we tried the DSO also. But from the ERP it is very difficult to generate such analytical ratios, which require data from many accounting periods. So we were not progressing well in this.

Waghmare: True, the ERP is more of a transaction processing tool. Trying to generate data intensive reports would not only be difficult, but also bring down the performance of the ERP.

MD: Yes, we realised that. So I brought in by BI expert and explained to him my predicament. He said that this should not be a problem at all. In just a couple of weeks he had two reports ready for me, the Inventory Turns report at Item level, and the DSO report at Customer Level. Both these reports are attached to my dashboard in different worksheets. In each of these reports, I could see a big picture and then also drill down to lower levels of detail. The BI expert had also created Inventory Turn and DSO Buckets, so that I could immediately filter out the bad performers.

MD (continuing): After that it was easy. I called in my factory manager and my credit control manager and showed them the reports. Both of them were amazed at the ease with which they could identify the items/customers, which were to be acted upon. Both said that they spent enormous amount of time in doing exactly the same things. They acted very quickly and in a month’s time, we started seeing the results.

Waghmare: So the important learning is that while in both the SBUs the senior team knew the problem, only in SBU B the problem was actionable. The executive team had a shortlist of items/customers to act on. While such information was not available in SBU A, and due to which they could not act on despite best intentions.

MD: Exactly!

Waghmare: Mr. Basu, I am consulting in various large organisations. In each organisation, I know for sure that the senior management team generally knows their overall data quite well.

They know the order book, the sales, dispatches, AR, collections. They may not have formal systems for this in place, but the informal network keeps them updated on critical numbers. However, what they don’t know is the intelligence which lies in the large data mountains. While they can hold the credit manager responsible for a high DSO, it is not enough. The credit manager needs tools to act on. In absence of the tools, he is helpless and the time goes by and the woes worsen.

Waghmare (continuing): I know of a company which had implemented a similar AR tool, and by consistent usage of the tool their AR ratio to Sales continuously reduced year on year.

MD: So, the key is in providing the right tools to the people in the front line.

Waghmare: Yes, providing the tools and its consistent usage will certainly bring in results.

Waghmare: So the key learning of the last three months’ experiment is that the deliverable of BI should not be just fancy reports; the deliverable of BI should be measurable business benefits!

MD: Well said, Mr. Waghmare and thank you for suggesting us to carry out both the horizontal and vertical approaches and learn from doing.

MD: So Basu, how would you like to proceed?

Basu: V…vertically, Sir!

Lata watches Mr. Basu come out of the MD’s cabin. His shoulders are stooped and he is wiping his forehead. She stifles a smile…

Observations
BI Initiatives must have well stated measurable business benefits. For example, AR DSO should reduce by 15 days, Inventory Turns should increase to 20, Period Closure should be 3 days etc.

BI initiatives also should be personally supervised by the CXO team to ensure that it receives the due importance and that the results get delivered.
 
Design Principles

• Design the BI reports in such a way that the users are able to see a big picture of the domain area.

• The BI reports should also enable the users to drill down to the problem areas.

• Highlight the problem areas by discretising the data, so that the problem areas are immediately identified and the user does not have to do the job of searching.

• The BI reports should also, wherever possible, show a trend of key KPIs so that the team is able to find out whether the situation is improving or not.

Sanjay Shah is the CEO of Pune-based Prosys Infotech

If you have an interesting article / experience / case study to share, please get in touch with us at [email protected]