The Biggest Dirty Little Secret in IT

I’ll get to the secret in just a minute, but first an observation about the cost of IT. Forrester has been conducting a cost of IT study for many years with the most recent results published in the 2013 IT Budget Planning Guide for CIOs. The report includes a chart of total IT spending as a percent of revenue by industry and company size. Cost as a percentage of revenue is a key performance indicator for IT efficiency as organizations increase in size. I first noticed a peculiarity in the data in the 2007 study and I was wondering if it had changed over the years – it hasn’t. The observation is this; for many industries, the cost of IT as a percent of revenue increases as organizations get larger. What is going on here? Whatever happened to “economies of scale?” Instead we seem to have “diseconomies of scale!”

Before I get to the secret (i.e. the cause of the phenomenon), let’s dig into the numbers a bit more. Since I don’t have permission to share details from the Forrester report, you’ll have to get the paper yourself to cross-check my findings. Figure 3 in the report shows the cost of IT in 21 industries and for each industry the cost for SMB (Small and Medium Businesses) and Enterprise (large) organizations. If you sort the list by cost, at the top of the list is Financial Services with 7.3% of revenue spent on IT, while at the bottom of the list is Consumer Products with 2.3% of revenue spent on IT. This makes sense since different industries are data-driven to different degrees. Banks don’t move money, they move data.

To continue, I calculated the difference in cost for SMB compared to Enterprise organizations and divided the list into the Top 11 and the Bottom 10. The results are startling!

Table 1: Average Cost of IT as a Percent of Revenue

In the Bottom 10 category, 9 of the industries have a lower IT cost percentage for Enterprise organizations compared to SMB (3.4% for SMB and 3.0% for Enterprise.) This represents roughly a 12% improvement in efficiency for large organizations and is what we would expect in terms of economies of scale.Table 1: Average Cost of IT as a Percent of Revenue

In the Top 11 category however, 9 of the industries have a higher IT cost percentage for Enterprise organizations compared to SMB (5.0% for SMB and 5.4% for Enterprise.) This represents an 8% decrease in efficiency for large organizations as they increase in scale. In other words, diseconomies of scale. The total difference then between the Top 11 and the Bottom 10 is 20%. In summary, as organizations with large IT groups become even larger, they are less efficient by 20% compared to organizations with smaller IT groups. The question is WHY?

Now let’s get to the dirty little secret – which isn’t so secret actually. The problem is the same as the cobblers’ children who have no shoes. IT spends a lot of time helping the organization to automate business processes, but spends almost nothing on automating its own IT processes. IT is largely a manual activity. Requirements are written by hand in MS-Office tools. Software is developed using powerful workstations and IDE’s, but still involves a developer essentially hand-coding the program. Handoffs between teams are through emails. Coordination of work is done in meetings. The list goes on.

The fact is that manual activities don’t scale linearly. Let’s take a simple example – a restaurant kitchen. With one chef, the output of prepared food is X. If you add a second chef, the output is pretty close to 2X. In fact, if the two chefs are smart, they might divide up the work so that each specializes on certain dishes or activities, and as a result becomes more productive overall. Even with 3 or 4 chefs you might see the productivity rise linearly as they start to develop a sort of “assembly line” or factory process. But what happens when you put 10 chefs, or 50, in the same kitchen? At some point they start getting in each other’s way, communications consume more and more time and energy, disagreements start to emerge, staff turnover becomes a productivity detractor, and so on.

The IT industry tried to address the skyrocketing costs by hiring lower cost staff (see my blog Outsourcing Doesn’t Make Sense Anymore) but in the end it is not sustainable. Large banks count the number of IT staff in the 10’s of thousands. If you truly want to scale IT, the solution is not to throw more people at the increasing complexity and the exponential growth in big data, the solution is to standardize the work process, development configurable re-useable components, and automate routine activities using appropriate technologies. In a scalable organization, you should be able to double work output by adding just a small fraction of staff.

So the biggest dirty little secret in IT is that IT doesn’t follow its own advice. It helps the business automate ordering, billing, design activities, manufacturing processes, and even entire supply chain management. Amazon may be the ultimate example by automating the order-to-cash cycle to the point where it happens with 1-click.  And it’s not just 1-click to purchase a book or pair of shoes, the Amazon Elastic Compute Cloud (EC2) also supports 1-click server provisioning or software deployment. It’s about time IT applied the same innovation to how we do work.

For example, one Informatica user had to rebuild 800 mappings in order to migrate an application from DB2 to Oracle. The Systems Integrator who bid on the project using the manual-based approach proposed a team of 60 people (mostly off-shore) for 4 months. The winning solution however was a highly automated approach using an Integration Factory approach and Informatica Platform that took one person one month to complete all 800 mappings!

You don’t need to wait. You can do this in your organization today with existing technology. If in doubt, come to Informatica World to talk to leading organizations who are doing this today. Hope to see you there!

This entry was posted in CIO, Data Integration Platform, Enterprise Data Management, Integration Competency Centers and tagged , , , , . Bookmark the permalink.

9 Responses to The Biggest Dirty Little Secret in IT

  1. Thomas Reichel says:

    Hi John, your blog is again spot on! I can’t agree more based on my own observations and experience. While many large organizations downscale the number of own IT employees by offshoring to low labor cost countries, the execution of business strategies is still hindered by lacking IT effectiveness and efficiency, mostly driven by stove-piped processes. Apart from the vendor hyped Big Data theme, I can see history repeating itself, hand-coding activities e.g. MapReduce programming using Java, Python and so on are growing rapidly, increasing complexity alongside the real Big Data business opportunities. Complexity will slow down any organization in achieving it’s business strategy in a rapidly changing ecosystem of economies, customers, competition, partners and so on…

  2. Marina Kerbel says:

    Interesting statistics and perspective. I think this situation will persist even if IT efficiency would have dramatically improved. There are too many external forces that continuously increase complexity of IT ecosystems (M&A, Cloud, Big Data, etc.). Short of impeding business goals, IT cannot stop it. With this increasing complexity, there is a tipping point after which IT efforts grow exponentially just to keep all pieces in sync and the systems stable.

  3. Jamie Wymer says:

    Great observations John. I think an even deeper analysis would be worthwhile, i.e., where do the companies rank in regards to ratio of IT spend to revenue vs. their EBIDTA performance?Or perhaps looking at their IT spend to revenue ratio and the comparitve revenue generated per employee is an even more meaningful insight into how much efficiency IT assets are delivering to the organization. Looking at other key indicators such as DSO and Inventory Turns (where pplicable) could also provide some insights on how well data is managed by the firms.
    My experience has been that the companies that invest the most in quality control and process automation generate the highest profit margins. Ultimately netting a higher ROI on every IT dollar spent.
    It would be very possible for a comany to show a higher IT investment ratio vs. revenue but substantially outperform their peers (having lower IT spend to revenue ratios) in profitability. The devil is always in the details but I believe this viewpoint of measuring IT operations for return on capital, documenting IT Process value streams and optimizing for quality, SMED capabilities and value add are the best approach to fulfilling the promise of best-in-class IT operations performance. Thanks for keeping a pulse on this winning management philosophy.

  4. John Schmidt says:

    Marina, I think what you are saying is that complexity and the rate of change reaches a point a where the ONLY solution is to manhandle. In other words, things are changing too fast for IT to standardize and automate processes.

    I agree that is a factor, but a big part of the problem is that IT introduces soooo much unnecessary complexity. Businesses these days are incredibly complex – necessarily so. But our siloed behaviors in IT and lack of discipline add a ton of extra complexity. So in the end, IT is its own worst enemy. But it’s our dirty little secret – so don’t tell anyone.

  5. John Schmidt says:

    Jamie, one of the challenges in doing benchmarking across organizations is the wide variation in accounting rules and how costs are tracked. That said, this topic really does deserve more in-depth study and analysis.

    I’m curious. How many organizations out there have:
    1. A well-defined application portfolio with an official list in a repository
    2. Where integration systems are one of the categories of enterprise systems
    3. Costs are tracked and reported on a regular basis against each application

    Does your IT group do that?

  6. Bob Kellum says:

    Everyone is focused on ROI. There has been a lot said about placing IT focus on the core business process of an organization. That means that there is an emphasis on creating automated processes for production of widgets, and all of the associated details.
    Placing more of your resources on the teeth make sense, but not taking care of the tail get’s messier over time. That is especially true of IT and its needs. Projects to automate IT are part of the tail, and don’t make for a good ROI story. You can drill down into the savings that you can realize, but with IT often understaffed, the savings will likely be absorbed. There is a large buffer effect created by the neglect of the IT process, and the technical debt that builds up over time.
    The only remedy is to focus on the update of IT resources and insuring that IT managers and executives are supported with the resources that they need. IT Managers also need to be more aggressive in getting away from the “fire management mode” that contributes greatly to the buffer effect that steals away the bang from the bucks put into the tail.
    It’s a difficult nut to crack, and can only be approached when executives make sure that there is sufficient (human, monetary, training, and expectation) margin to realize economies that are typically reserved for the higher visibility items. Any function that operates at “bare bones” resources will find it very difficult to climb out of fire management.

  7. Wes Anderson says:

    John, nice to see some numbers to back up what a lot of us at big companies see. I think you hit the nail on the head when you say “It’s about time IT applied the same innovation to how we do work.”
    In my opinion, the decentralization of data integration compounds the problem as large companies with a wide variety of technology teams are given the flexibility to choose their own tools and paths for their particular projects and applications. If there is no common strategy, there is very little economy of scale or opportunities to reuse processes, code or lessons learned.

  8. Nic Harvard says:

    When you have a small/micro operation, the InfoTech team is usually small enough to fit in one office.

    When a business problem arrises, all the prior learning and experience is at their fingertips, and they will naturally re-use as much “stuff” as possible to save effort.

    As soon as one gets teams of hundreds serving organisations which turn over billions per annum, one runs into this “too many cooks spoil the broth” thing.

    The normal reaction is to apply control, governance, and portfolio management.

    But this is usually done by non-technical people, and (amazingly!) forms yet another silo and barrier.

    My calculations are that over 50% of most IT-related costs in medium-large projects are frictional – project managers and other “boilerplate”…

  9. John Schmidt says:

    Nic, your finding that 50% of costs in medium-large projects are frictional (or in lean terminology non-value add) is consistent with my finding that 50% of integration infrastructure in large enterprises is unnecessary complexity. The percentage is less for organizations that have a strong ICC and a relentless push to simplify and rationalize systems and technology. While we shouldn’t feel good about this much waste, it does reinforce how big the opportunity is if we could eliminate it. Which is of course why I am such a strong proponent of lean.

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>