POINT Sunset is looming – Will Historical Data be History without Pre-emptive Steps?

Barclays POINT Historical Data Migration

Among the myriad challenges facing companies transitioning from Barclays POINT is how to migrate years of historical analytics, and perhaps performance, to another platform.  Competitors vying to win POINT replacement business are doing everything they can to reassure companies that, with their solution, the historical data migration is going to be smooth and easy.

But as an industry, how many of us can look back at any project of this scope and complexity and remember, “oh yeah, that went a lot smoother than I thought it would?”  And so it will go with the transition of historical data from POINT – however fraught with challenges it appears in the preliminary analysis, you can rest assured there are more than a few unpleasant surprises in store.

For companies who want to avoid any surprises, there are three major challenges they should address before they begin their Barclays POINT historical data migration.

Three Major Challenges

1) Reconciliation of Historical Holdings – Re-writing history

One of the biggest issues companies will have to face individually is reconciliation.  Of course there are lots of reassurances out there from vendors that “moving the data over” won’t be a challenge, but nobody should rest easy on this account.  Firstly, if you’re moving over years of historical holdings or transactions-based data, you’re literally going to be re-writing history.

POINT is well known for the sophistication of its risk and attribution models, which come with a hefty price in terms of the detailed nature of the data it requires.  It’s simply not possible to feed all this data into a completely different model and expect the same results.

Even Bloomberg, who have been very confident that they can help companies migrate data (and they should be, since they’re inheriting the POINT architecture), concede that it is their client’s responsibility for reconciliation, analysis, remediation & reload once the data is moved.  To the extent that companies want to preserve their POINT history, this needs to be a primary transitional concern.

2) Volume of Historical Data Overwritten – Doing it correctly the first time

Secondly, the volume of historical data itself raises issues, and stands to amplify the reconciliation matter.  For companies that intend to overwrite analytics in the new system, there has to be an organized process and they must have the tools in place to ensure this is done correctly the first time.

Consider for a second just a handful of issues surrounding overriding analytics for more than fifty portfolios, containing perhaps tens of thousands of assets, every day for eight or more years…now compound these challenges with the fact that your source for analytics overrides is coming off-line within a year, which means the historical transition probably needs to be complete within the next six-months at the latest.

Just the scope of the project means it’s likely several weeks in execution assuming you have people to dedicate to the task.  If you nest all of this in the transition timeline, it becomes obvious that you cannot afford to take an iterative approach to overriding analytics.  It must be done correctly the first time, and because the results will not reproducible in the new environment reconciliation is even more critical.

3) Perils of ‘Do It Later’?

Lastly, there are a number of companies whose approach involves pulling everything possible out of POINT and storing it somewhere until after a new system is in place and then, without the pressure of the transition timeline, deal with the history in a methodical way.  At first glance this seems like a low-risk, sober approach to handling the history – and it has its advantages.  It also, however, presents some unique challenges.

With the “do it later” approach

• A data warehouse needs to be built, or space and an organizational structure for the data needs to be created within an existing environment.
• Samples of the historical loads should be tested in the new environment, to avoid the significant risk of discovering that they are missing a critical element of the history…too late.

And if you’re going to run tests on all these loads prior to archival, it begs the question of how much work you’re really saving yourself since there is disproportionately more time involved in doing the first load (mappings, file configuration issues, etc.) than there is in the rest of the history.

As the POINT sunset approaches, keep in mind that there are two sets of problems out there related to historical data.

• There is the well-known, universally experienced variety such as the limits on historical benchmark data. With these generalized challenges, a best practice will emerge and companies will be forgiven for doing as well as they can with an industry wide problem.
• The less forgiving types are the problems unique to each company, and which will challenge managers outside of the relatively safe environment of an industry-wide evolving best practice.

Managers involved in the problem of POINT historical data must work to aggressively identify and bring tools and resources to bear against these problems, with the full expectation that several complications they cannot yet foresee will arise in the process.  There are tools available.

The clock is ticking.

Ensure business continuity in your POINT migration with IKINDI KTier™ Contact us

This post, originally published on 11/1/16, has been updated and expanded and republished here.