Migration from Deski to Web Intelligence

As you probably know by now, with the release of SAP BusinessObjects Business Intelligence 4.0 (BI4), Desktop Intelligence (Deski) is officially retired. Of course, we’ve known that this day would come for nearly ten years, and some people started early in migrating their Deski documents to Web Intelligence (Webi). But many also delayed until the last minute, waiting for Webi to mature a little more. But the time for waiting is over, and it’s now time to put Deski behind us. So, to that end, I’d like to share some experience in converting those Deski documents to Webi.

The Right Tool for the Right Job

Most of your Deski documents can be converted to Webi with the help of the Report Conversion Tool. If you haven’t used this tool yet, try it out with one of your Deski Documents, and see how it does. I’ve used it many times, with varying degrees of success. Sometimes it converts complex documents, with no issues, much to my surprise. Sometimes it fails to convert very simple documents, much to my dismay. 😦

So, my first piece of advice is, if it fails to convert a simple document, don’t waste time trying figure out why. Simply rebuild the document, from scratch, in Webi. I hate to admit it, but I learned this lesson the hard way.

I hate failure, and am usually not willing to give up easily. One time, I ran a Deski document through the Report Conversion Tool (RCT), and it failed. I couldn’t figure out why. The document had one query, one table, and one report, with only one simple variable. I spent two hours trying to figure out why it wouldn’t convert successfully. Then it dawned on me. I could rebuild it in Webi in five minutes. It wasn’t worth the time.

So, for simple reports, try it once with the RCT. If it fails, simply rebuild the document.

Did it Work?

The RCT has three potential outcomes for each document:

  • Converted Successfully
  • Partially Converted
  • Not Converted

If the conversion is successful, compare the Deski version with the Webi version. Make sure the formatting is correct. Many times, it will need some touching up. And make sure the data is the same. Not all functions work exactly the same, so you may need to modify the formulas in some variables to get the correct results.

If the document is partially converted, compare the two, and see what didn’t convert. For example, if your Deski document has any macros in it, they won’t be converted, as Webi doesn’t support that feature. You may need to get creative to find another way to get the same results.

If the document is not converted at all, the RCT will tell you why. Then you get to decide if it’s easier to try to solve the problem, or simply rebuild the document.

Freehand SQL

I’ve struggled a long time with this. Best practices tell us that we should never be using Freehand SQL to build reports. It defeats the purpose of using BusinessObjects. It is very difficult to maintain. And it often violates corporate security policy. But that hasn’t stopped people from building these reports. Sometimes it’s just easier to write the SQL by hand, rather than modify the universe to accommodate the report’s requirements. I get that.

So what do we do with these documents? Well, the good news is that you can use the RCT to convert them to Webi documents. However, since Webi doesn’t support Freehand SQL, how does it work? The RCT will convert the document by generating a universe with a derived table in it. The Freehand SQL from the Deski document will become the derived table in the universe. If the document has multiple Freehand SQL data providers, each will become an individual derived table in the same universe.

In fact, one such universe will be generated for all Deski documents that use the same database connection, and have Freehand SQL data providers.

But don’t leave it there. Once converted, start migrating these new Webi documents to real production universes. Simply change the universe for each query in the Webi document from the auto-generated universe, to your production universes. Finally, after you have all the new Webi documents attached to production universes, you can delete the auto-generated universe.

By the way, this also applies to Webi queries that were built with universes, but then the SQL was edited.

Marketing and Political Unrest

One of the most challenging tasks in converting from Deski to Webi is the fear of change coming from the user community. Let’s face it, once you get good at a tool, you don’t want to be forced to change to another tool. So, we need to make this as easy as possible for the users. This starts with two very important jobs:

  • Management Support: You have to have management behind this decision. If they aren’t, you can find yourself fighting political battles, and not always winning.
  • Marketing: You need a campaign to “sell” Webi to the user community. You want them on your side.

Both of these are interconnected. If the user community screams loud enough, management will not support the project. So, get a few key, influential users, who are respected throughout the organization, involved in the project. They will be your marketing evangelists among the rest of the community. They can help get the rest of users excited about this change.

Management will support the project based on the cost savings realized from supporting one tool, rather than two tools (One of which is reaching “End of Life”).

Are You On Board?

I’ve seen many posts on BOB from IT folks complaining of this change. Some have threatened to leave BusinessObjects and move to another tool. And if that decision makes good business sense, then do it. But don’t change tool sets simply because BusinessObjects, and SAP, have decided to drop Deski.

I grew up on Deski. I have spent hundreds, if not thousands, of hours developing reports in Deski. I become extremely competent with the tool, and even taught the classes for many years. But over the last five years, I’ve used it very little. I’ve focused my attention on Webi, and I’m glad I did. I like Webi very much. And I don’t miss all the quirks and bugs (Remember the Unhandled Exception Error?) in Deski. Yes, I know, Webi isn’t perfect. But it’s a great tool. So, make sure that you’re learning it, and excited about it, before you begin your conversion project.


The bottom line is that Deski is now retired. Honestly, I’m happy about that. As much as I loved that tool, it’s much easier to focus on one less tool in the tool belt. So, I hope this article has given you some useful tips converting those Desktop Intelligence documents to Web Intelligence.

What is your experience? Do you have any additional tips you could share?


7 Responses to Migration from Deski to Web Intelligence

  1. Nice article, I fully agree.

    I sometimes train users to make the shift to Webi, and many of them are crazy about the missing functionalities. That’s better with SAP BI 4, but it seems that grouping and some criteria in query filtering are still missing.

    In XI3 at least, page formatting must be redone. Some differences with cumulative sums or graph types make a rewrite necessary. The stability of Webi is not good (much worse than Deski in my experience). And so on.

    After years, I know most of the traps and how to work around them, but even IT users cannot expect to know all of this.

    You’re right to say that you need the management’s backing, because the migration is for most users a heavy useless task. They don’t see the benefits of going the Webi way, at least at first, and wonder why so many functionalities are gone. Some novelties are nice, but it does not reduce their conversion workload. I’ve got many customers locked with Deski, they will suffer when forced to migrate to the Deski-less XI4. Or they will pay me and others companies to convert all of this for them, a painful and uninteresting work that sucks budgets for more interesting tasks.

    I do not regret Deski and its own limitations, but BO/SAP made a big mistake there. They’ve lost years to re-writing their main product without providing an painless migration path. During this time, competition has got better.

  2. I agree with your sentiments on Freehand SQL. I’m a little bothered about SAP considering adding the feature to Web Intelligence. I guess I might be OK with it if the tool provided more than just a text box to paste SQL. More incentive to buy Information Steward, assuming the Metadata Management features trace freehand SQL?

  3. After many years of taking a firm stance against Freehand-SQL I have reversed my opinion, I believe Freehand-SQL should be available. Not only report develpopers, but also some business analysts have SQL skills, and if access to creation of Freehand-SQL based data providers can be controlled AND audited (dependent objects) easily, I believe the advantages of this flexibility easily outweighs the disadvantages.

  4. Jon Fortner says:

    We have 7000 DeskI reports, so this effort will be very involved. We plan to first do a cleanup of unused reports using Auditor and inventory data. Once there, we are considering our options for VBA Macros. They were mostly used to loop through a list of Suppliers/parts/etc. run the report and send it via E-Mail. We think you can use the WebI client and an Excel file to get the list, then use a Publication and WebI report to do the rest. Anyone have other suggestions?

  5. Sreeni says:

    Very good Research and on Deski.. Very Usefull Info.. I agree with you.

  6. swati says:

    I have one question here, when i have run deski it is working fine but in webi it is giving session time out error , with in seconds.
    Can anyone faced this issue?

    please give the solution if anyone knows.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: