“Pay attention to what I say, not what I do…”
Let’s return to the bonus chart slipped in at the end of the recent post @DeutscheBank: Three-Card Monte and Other Confidence Games (For maximum context and extra credit, you can pick up the thread about Deutsche Bank from the beginning in March 2016 here).
Bottom line: We are fascinated with the idea of detecting the impacts of the IT outsourcing deal that DB and HPE entered into in early 2015. The main questions that keep coming to mind are these: Is this deal a template for other large banks? Does it save money? Or, is its value to be found in other metrics, like enhanced performance or boosts in innovation? And, based on the level of transparency provided by DB (which is one of our best bank models), can we even detect the impacts of this arrangement?
The answer to this last question is what brings us to this post…
Right up front, we can say, yes – generically, speaking – outsourcing technical infrastructure to specialists is an appropriate, even necessary, move for the current era. With very few exceptions – mainly related to the highest-performance use cases – no large entity, like a global bank, should be managing their own infrastructures anymore. The available solutions have become mature enough and functionality continues to improve. Plus, banking and other large financial services firms need to gain comfort with idea of a lengthening supply chain, where technical infrastructure is among the first core components that needs to be pealed away from the proprietary mindset of the prior era.
Now, whether outsourcing to HPE is the best call for DB – or any other large bank – is really not up for debate here. Maybe it is, maybe it isn’t. It’s the supply chain configuration itself that attracts our curiosity, now and for the foreseeable future.
With these points as backdrop, in the exhibit below, Alphacution starts by illustrating DB’s quarterly technology costs per employee for the period Q1-2011 to Q1-2018. Recall that the DB-HPE IT outsourcing (ITO) deal was announced in Q1-2015, so they are now 3 years in.
In this chart, we display quarterly technology costs for the past 29 quarters. On top of this, we construct two trendlines: one for the entire data sample and a second for the 12-quarter period since Q1-2015. The purpose of this bit of detailing is to detect any change in spending trajectory since the ITO deal was consummated.
Setting a more detailed quantitative explanation aside for now, these two trendlines basically suggest that there is not a material change in the quarterly run rate for IT costs yet. If anything, they appear to have become more volatile. This is the best case reading, so far.
Worst case, the trendline since Q1-2015 (red) actually shows a slightly higher cost trajectory (with HPE) than the trendline for the entire data sample (blue). In any case, the impact that HPE is having on these costs, so far (3 years into a 10-year deal) is not obvious from these data.
So where’s the beef? Recent disclosures show that, if there is any material savings to be found, it’s in labor cost savings – which is where the biggest lever for savings usually resides in any case.
Here’s the setup: Like a vast majority of firms, labor costs are the largest component of operating expenses. DB is no different, with recent staff costs consuming about 50% of operating expenses (which has been on a downward trajectory from a high of nearly 65% in 2006). The chart below is a record of this broad trend…
Meanwhile, many banks like DB have been reorganizing technology, operations and other enterprise support personnel into a “corporate center” group (which is a key component of the platform development phenomenon we will be talking more about soon). In the case of DB, roughly 1/3 of total headcount of 97,130 – or, 32,131 as of Q1-2018 – is now designated as part of the corporate center. This is where technical infrastructure personnel live (see below).
This group is where to target the greatest amount of cost savings from improved management of technical infrastructure (and certain enterprise-scale applications).
There are three basic ways to save here: 1) minimize headcount, 2) minimize compensation, or 3) both. However, there is only ONE way to save on technology while preserving the possibility of improved technical performance for the future: Optimize headcount, skills mix and compensation (which is Alphacution’s preferred long-term strategy).
Recall from much of our prior writing: Spending on technology is not the priority. What a company receives from its investment in technology – the “return on technology” (RoT) – is much more important than absolute spend on technology, and that’s what brings “performance” – which we often measure as revenue per employee (RPE) relative to tech spending per employee (TCO/e), otherwise known by our proprietary analytic, “T-Spread” – into the discussion.
So, returning to our review of this ITO deal, the near-term objectives seem to be much more existential in nature. DB’s “performance” as measured by RPE has been under pressure for several years, since the global financial crisis (GFC), with consistent declines in group RPE for most of the past several years (see chart below).
And, these declines are pervasive, impacting the most high-performance groups – namely, Asset Management and Corporate & Investment Banking (CIB) – most of all.
By these and other measures, DB is in a state of turmoil, and therefore, it cannot afford the luxury of optimizing for performance. It must contain costs wherever possible, first.
This focus brings us – finally – to where the ITO deal may have yielded some progress. In the chart below, Alphacution illustrates the average annualized compensation for the corporate center segment (aka – support personnel) over the past 10 quarters.
So, when we translate that level of per-employee savings across the entire corporate center group – and assume that some component of this savings is occurring as a result of the ITO deal and related maneuvers to rationalize tech spending – we arrive at a meaningful level of aggregate savings: Over the past seven quarters, Alphacution estimates that DB has saved as much as $1 billion compensation expense for support and related personnel. And, given the timing, it’s not much of a stretch to lay the cause of some of this progress at the feet of HPE.
In short, while DB may be spending as much or more today on the boxes and other technical trimmings as they where before the HPE deal, Alphacution estimates that the cost of the staffing to manage all of it was down 10% from 2015 to 2016, and another 14.2% from 2016 to 2017 – for total 2-year savings of $923 million.
In closing, and though there is good evidence that the ITO deal has its merits, where we find our curiosity remaining unsatisfied is in the area of performance and innovation. We can see where this new arrangement has satisfied near-term cost pains, but it remains unclear if it also sets up the bank for better performance on a longer timeline. Not to be too critical, we believe that the act of setting up a centralized platform, as DB appears to be doing here, is a precursor to those kinds of fireworks down the road.
Certainly, we’d like to see new levels of performance show up in the RPE analytics (or, our proprietary T-Spread analytic), where a given level of performance is being achieved with a new (digital era) mix of technology and human capital.
That said, we know that innovation starts small; typically involves its share of experimentation and failure; and, can take a long time to be confirmed by the numbers. We also know that large institutions – especially those born in the “analog era” – have a notoriously difficult time getting out of their own way, and moreover, that some never do…
Alphacution will continue to dig for clues of these types of transformations, and report back when we make discoveries worth sharing…
As always, if you value this work: Like it, share it, comment on it – or discuss amongst yourselves – and then send us firstname.lastname@example.org.
As our “feedback loop” becomes more vibrant – given input from clients and other members of our network, especially around new questions to be answered – the value of this work will accelerate.
Don’t be shy…