Thoughts from TSAM (Part 2)

Following my most recent blog on the panel discussion held at TSAM in London recently, I thought I would add some further notes on the discussion. We had talked about buzzwords used and then about data governance… The theme of the discussion then switched to the risks you’re exposed to from miscommunication of information or data – the commentary is really as you would expect:

• Fines from regulators

• …which can lead to brand damage

• …which can lead to loss of clients and mandates

• …which does lead to outflows

• …which does directly impact your bottom line

Of course, the point was made that you do not need be fined by the regulator to incur the spectre of spectacular outflows – poor data quality in client communications is enough to trigger this alone.

I related a specific story I had been told by a director of institutional sales at a prospect I met in not too distant past, who earlier that month had got through the RFP process for a serious eleven figure mandate, which would generate many tens of millions in fees. So having got through the RFP process, this manager clearly had the right risk/performance figures to meet the minimum hurdle for inclusion in the beauty parade process. The deal was lost though on one critical point – the data presented at the beauty parade on the sales deck was completely at odds with the strategy performance as listed in the RFP response, and yes, they did not win the mandate. If you are handing someone billions and billions to manage you need to build a relationship based on trust and transparency and having inconsistent/inaccurate data leads to total breakdown in trust.

The topic switched again at this point to how can we get IT and business working together more effectively on data management projects. This topic generated lots of interesting viewpoints, which I have summarized here in bullet form:

• Business often gets involved too late – something specific to IT led projects

• There is general consensus that business-led projects are more successful mainly as the requirements are understood earlier in the process

• The project analysts and the project manager need to have strong business domain expertise with a really good understanding of technology to bridge the gaps between two teams

• It is not easy to find analysts with good understanding of IT and business – panelists agreed that the more successful people are those who start in IT and move over to business side.

At this point the discussion started to wrap up after a few questions from the audience and each panelist gave their final thoughts on overcoming the challenges. My own thoughts were that the driver of data management projects is changing, it is no longer fear of fines, it is sales and distribution demanding timely and accurate data. Another viewpoint was that we have to do a better job to remove artificial differences between IT knowledge and business knowledge, greater effort is required to try and get people to understand each other’s point of view. As data management projects are getting more complex, clear objectives and accountability are key success factors, we have to get the right stakeholders involved and use the right language. Finally, one of the panelists said we should not see data governance as a cost!

It was a great session, and I really enjoyed sharing views with the other participants on the panel. I look forward to the next one in New York on May 16th.

Thoughts from TSAM (Part 1)

I attended Osney Media’s TSAM conference last week in London – I always enjoy the event, it brings together senior level industry practitioners for thought-provoking discussion and debate. I participated in the panel discussion in the data management stream entitled “Overcoming the challenge of communication issues within data management”. My co-panellists were Shannon Walker, Business Architect – Finance Change, Deutsche Bank, Arun Sarwal, CEO Investment Management Solutions at DST Global Solutions and David Renn, Head of the Data Management practice at Citisoft. The discussion was moderated by Gary Pringle, Vice President, J.P. Morgan Asset Management.

There were a few areas covered in the discussion. First-up, we talked about buzzwords we use in data management … the theme quickly turned into – buzzwords that cause most confusion. There was fairly unanimous agreement that the expression “meta-data” was the most misunderstood buzzword being bandied about and the one that causes most confusion – so meta data is data about data – everyone clear on that – but then it splits into meta-content (data about the data content) and meta-structure (data about the structure and data container). Other buzz words the panel agreed caused confusion:

• “Data-Warehouse” versus “Data-Mart” versus “Data-Store” versus “Data-Hub”; then, what about “silo”, “operational data store”, “historical data stores”, “as-of  / effective on data stores”.

One thing the panel unanimously agreed was that data warehousing is a toxic term at C-level! Maybe this is why so many different other expressions have evolved to describe data warehouse like systems….

• Managed Data Services – a lot of vendors are in the market with data management in the cloud with value-added “managed data services” (yes I am one of them…..) but each vendor means different things when they bandy that term about – where the managed data service starts and ends on a vendor by vendor basis can be extreme..

The theme of the discussion then changed to ‘Governance’ and how strong governance could lead to lower risk – as expected no one had a contrary argument – but the views of the panel were quite clear on one point – many firms talk the talk about governance, they have appointed their Chief Data Officer or Tsar, but in general there is not enough bite or budget connected to the role to allow it operate efficiently. My own views are fairly black and white on this point – too many firms pay lip service to the application of an efficient stewardship function that actually walks the walk that the governance speaks to – governance (strategy) without stewardship (tactical execution) is a pointless window dressing exercise.

There were many comments in earlier discussions that mandates now routinely ask in-depth questions on the governance structures in place, expect these due diligences to extend into examination of the tactical application of the governance to expose the level of stewardship in the firm and its relative effectiveness in carrying out the governance remit.

My own views on governance and the reasons it remains an area for investment are simple

• Governance with effective stewardship will drive costs lower and reduce risk exposure
• It positions firms to take control of their own data
• There are demands in the front-office from distribution to broaden the depth and breadth of data and demands for a more agile approach to product data publication cycles
• Data is the oil in the sales engine – without oil the engine cannot run, with dirty oil the engine will run but eventually seize up.
• The increased demands for data quality, depth and breadth is driving requirements for scale in compliance and audit
• The cost of managing data is going up and firms not set to scale in this area are being left behind.

I will end this section of the blog here, as it’s getting quite long and I’ll post another one about the rest of the discussion in a few days time.

What risks are at play when you use a third party data provider to populate your data on your own website?

In this age of renewed focus on cost savings, I have noticed quite a few firms are increasingly using third party data provider data for their own products on their own websites as a way of reducing their internal data management costs. At face value, it seems like a reasonable way to drive down cost for what can be a very serious cost driver in any asset management firm.

However, you can think about it in another way. What message does this give to the market, to the potential investor? Does it say that you are incapable of managing your own data, or even that a third-party data provider has a better grip on your data that you do? What does this say about your firm?

If the data provider errs and promotes inaccurate information about your product on your website, sure they may take a hit – maybe a month’s fee in SLA credits, but who takes the true financial burden, who takes the reputational hit, who deals with the regulators who arrive on site for a multi-week due diligence audit?

Also – people may make investment decisions based on incorrect or out of date information – customers could choose to withdraw their investment, and new prospects may decide to invest elsewhere. If it turns out, that your funds were mis-represented in public, you will suffer damage to your reputation and to your brand and you might even have to take corrective action if any investor loses money due to errors in the information that was provided.

It all comes back to caveat emptor – buyer beware – you get what you pay for.

Where are you on the Data Governance evolution scale?

This post was recently published on TABBFORUM (21st February)

Just about every asset management firm now claims to have a formal data governance process in place – in fact if firms aren’t saying this we should be worried indeed. So while all data quality management governance processes may be created as equals, they are very rarely at the same point of evolution – with the bottom of the evolution scale being a million miles from the top-end fully evolved processes.

In the diagram below you can see that processes in the early stages of evolution start as chaotic processes, with often low levels of standards and formal operating procedures, with very little sign of an obvious ‘master plan’  or strategy.

The Data Governance Evolution Scale

In order to move to the next stage on the evolution scale, you need to establish standards, you need formal operating procedures for data stewards such that a semblance of an operating data quality management process starts to take effect with a strategy and master plan identified and communicated to the applicable stakeholders.

The mid-point in the evolution scale is achieved when the process can be accurately described as defined – that is where you have identified key performance indicators that show the health of your process, where you have documented artefacts such as a data dictionary and rules dictionary, where you can show you have stewardship operating across the breadth of the data creation to data consumption processes, with applicable technology frameworks in place to support stewardship of the governance with key processes like root cause analysis being tracked and measured.

Very few firms have moved past the ‘defined’ stage in the evolution process. Getting to the next stage ‘Pro-active’ requires serious attention and investment and very often monumental cultural re-alignment. Pro-active governance is achieved when you can demonstrate a cast iron continuous improvement cycle, with error feedback loops constantly leading to process improvement – very much in the model of six sigma – in fact many firms who have attained this level, do so under the auspices of an investment in ISO9000 or Six Sigma. At this point of evolution, the firm is applying automation across the board to root out the manual human errors that plague many firms today. Key to the approach is a unified governance approach to how data is managed across all of the data silos in the firm.

Finally, you have reached the nirvana point of evolution when your data quality governance has become what many refer to as ‘Pre-dictive’. At this point of evolution, not only is the process fully automated, it also has a fully demonstrable audit trail that fosters accountability and ownership. The top-down strategy is fully in tune with the bottom-up application of the strategy, with complete cultural alignment across the breadth of the firm, effectively with the people, the process and technology all working in harmony. At this point, your process feedback loops are fine tuning, rather than fixing.

Tribal conflict on the investment plains

One of the key trends now with the global asset management community is the redistribution of products from one region into another – take for example, the popularity of BRIC and emerging markets funds in the US and Europe. Similarly, in Asia you have very high demand for equity funds from the G7 regions, and investment grade bond funds from those countries lucky enough to retain their AAA ratings. Global firms are increasingly co-locating their investment management teams in the regions where the investment is being placed. The middle office support for these teams is also increasingly being co-located with the same teams.

The problems start when the fund is sold in another region, quite often the local sales/distribution team takes the core investment product data from the local team and applies their own slant to the information – this application of regional slants to data coming from the region of investment can lead to very serious consequences, which can often erupt in tribal conflict between the regional division producing the product and the regional division selling the product.

Simple things like re-classifying terms such as ‘Real Estate’ (US lingua) to ‘Property’ (UK lingua) can seem straightforward, but when you have one region that takes a security classified as ‘Asset Backed Security’  and changes this to ‘Cash or Cash Equivalent’, problems can emerge. This is a simple example of course and one that very few firms will make again, but there is unlimited scope for misunderstanding and resulting misclassification of data when you have one team trying to interpret what another team means.

The only way to solve this is for global firms to have global governance and stewardship for their investment product data. The distributed / decentralized model for governance which exists in many global firms today will only lead to continued conflicts between their regional centres, and in turn expose their firm to specific reputational, regulatory and financial risks.