IT Cost Savings... I can't stand this topic

A huge part of me despise talking about this topic because there’s a tendency for misinterpretation. Secondly, I am a staunch believer that you can’t cut yourself to profitability or growth. Money needs to be spent and value derived from the investment. 

Hopefully by unlocking value from wastage, the results can be redirected for both growth and innovation.

So this post shall focus on cutting away wastages and excesses in the IT organization. The language is harsh as wastage is unacceptable.

Signs that there’s a ton of fat to be cut - roughly 50% savings at least in a recent case that I went through:-

Missing semi to real time Configuration Management Database

A Configuration Management database tells you exactly which assets are in used, what are they for and to what degree of utilization.

Without this, the following wastage are highly common:-

• Underutilized hardware 
• Buying more hardware or subscribing to more license than you'll ever need 
• Continuously paying for items that don’t even exist

Missing Service Level and Vendor Performance Management practice or reporting, or just shoddy reporting quality

Usually translates to vendors and service providers running circles around the IT organization without anyone knowing what’s happening or, for political reasons, the vendors are able to get away with it

No Demand Management and Program Portfolio Management practice where inbound IT projects are tracked throughout their request and implementation life cycle

Which is typical for highly political organizations where the most influential department head gets to request any manner of new projects and de-prioritize one that’s already in flight leading to stranded and scuttled resources. I’ve seen first hand projects cancelled because a department head wanted to impress the CEO with a personalized app as the CEO’s birthday is coming… not kidding.

How do you even start? - Milestone #1 - Low Hanging Fruits

Begin with the invoices

How much money is being paid out to whom. Where’s the big ticket items and what are the frequency of payments? Dodginess abounds when there’s a string of invoices that’s magically priced below a certain level of authority paid in rapid succession to avoid scrutiny. 

Market benchmarks are helpful here to also understand whether it is more than the market average. 

It can however, be a challenge as invoices will be rather cryptic, and when there’s poor finance governance as well, be rather difficult to trace the invoices to their original contracts.

Trace the invoices to actual contracts and scope of works

Because fat also means that the invoice payments will not commensurate with the scope of works. For services, it typically means the contractors or personnel supplied are of way lower quality or practically unable to service the entity based on the level of capability contracted. Some don’t even exist or come to the office only once a year for a chit chat to book their engagement.

Another sure fire sign are ambiguous contracts with little to no tangible measures and deliverable. The company is pretty much screwed and will need to undergo negotiation as even the grounds and stipulations for early termination can be vague or scarier, do not exist. 

Some organizations have more form than function when the Vendor Management unit actually don't understand the scope of work or legal contract - not kidding again. Reason cited, we're really "procurement" people and not "IT"...

Find the underlying IT assets behind the contract (Location rental, hardware and software as well as subscriptions e.g. clouds or networking, telephony etc.)

Cause, you’d be surprised to see that they either no longer exist, stored in a room somewhere, severely underutilized as the software was practically forgotten after it was commissioned, or far fewer in items compared to what was originally contracted.

There was a project back in 2007 where ~12,000 applications had to be culled and consolidated. We ended up just shutting them down based on the last timestamp of access - no one made a peep except for a handful of incident calls. That brought the numbers down to ~3,000; and after a bunch of emails sent out and calls as well as discussions, it went to ~1,500 before the real consolidation effort actually started.

Find who in the company and customers are actually using these assets and should be signing off on the payments

Interestingly enough, in poorly governed entities, there’s not even a concept of an Application or Data Owner. Or even if they do, the person potentially has already left the company or it’s the executive that’s left remaining in the department. Finance pays the vendors like robots, and everyone gets their paycheck at the end the month. 

If the stakeholders no longer exists, then the solution is pointless.

Milestone 2 begins after you’ve covered all the bases above. As in performing the tasks, you would have the benefit of a birds eye and worm view of the entire IT landscape 

Milestone #2 - Tactical Assessments and Activities

Given sufficient resources these activities can be performed in parallel.

Consolidation of under utilized assets 

Databases are the easiest, just merged all them e.g. Oracle, SQL Server and DB2 databases together as their individual pieces. Followed by secondary migration of code to the cheapest RDBMS with the best support that you can find. It’ll take several months to several quarters depending on how horribly managed the place is. Consolidating multiple underutilized databases into a few is not rocket science, and is the easiest way to lower cost of licensing based on CPU or instance.

Followed by consolidation of operating system instances into virtualized or containerized environments. Similar time frame roughly multiplied by 1.5X - 3X

Migration to the Cloud 

Assets that are primed for the cloud are customer facing solution with high evolutionary rate of change. Unfortunately, not many companies move as fast. Having spent a ton of money for the migration, the customer facing apps practically stays the same for the next 3-5 years and gets scuttled for a new app that may look nicer but have only 50% of the features. The people related issues need to be changed first here.

Second category of assets made for the cloud are systems that are periodically elastic by nature. E.g. come Christmas, the website will be inundated with hits as the Chief Marketing Officer thought it was a great idea to have e-vouchers on the website and promote it on prime time national TV. Besides that, no one ever visits your website... 

Identifying Core Assets versus Commodity Assets

Commodity assets = stuff with a ton of suppliers and options

Put up RFPs for commodity items like laptops, desktops, printers, office IT equipment including management and support. There’s no reason why this needs to be performed in-house other than there’s a lot of emails that needs to be ‘hidden’. And NO, company issued phones should not be the latest iPhone. Just work out a deal with the many telcos that we have out there with a phone package.

RFP for price agreements that’s transparent for everyone to see. One thing I admire about US government agencies is the notification and uploads for all to see, price agreements for software and hardware post RFP. 

Fake/Red Herring Core Assets 

Core Assets = IT capabilities that are core to business success, and contains game changing intellectual property. 

But let’s get serious, if you’re a bank with the same fixed deposit, loans and savings products with any other bank except changes in interest rates - the core banking system is unfortunately NOT a core asset. So, do find a cheaper one when the organization lacks the ability to innovate.

These projects unfortunately, takes forever to complete if they’re companies that had to go through Milestone #1. 

Other fake core assets example include a mobile app for basic banking services. Again, payments, checking savings account, transfers are a dime in a done. If there’s no differentiator to the next guy then keep it simple and economical - get something off the shelf that can be quickly re-branded...

...stop building mediocrity!

Real Core Assets (Examples)

• An AI investment that you’ve trained for the last 2 months to answer customer online chat queries and accuracy rates have hit 96%.
• A trading bot that’s been fine tuned and returning decent earnings on the specific portfolio that the company invests in, except you don’t have to pay the bot 15 months bonus
• A performing operational organization with seriously fast turnaround time and level of accountability - don’t disrupt them and do keep-em happy

These investments needs to be continuously enhanced and guarded carefully. You wouldn’t want your competitors to duplicate the hard earned effort by stealing your banking product’s taxonomy in the AI example.

Milestone #3 - Strategic Assessment of IT 

Answering larger existential and operating model questions.

Outsource versus In-source?

• Should we create an IT subsidiary as a profit center? Would getting a stake of the pie make the entity perform better or be a direct award parasite to the group thus consciously raising prices to the point that the group changes their mind…  
• Should we outsource it to a vendor that's willing to put some skin in the game where performance of the company is tied to performance of IT. Rather popular in the early 2000s but largely died-out as organizations lack governance saviness to maintain this relationship for the long term.
• What about building an entire army of programmers and technical team - works great when there’s strong leadership to have them work together as a TEAM versus siloes of technology pillars that don’t talk to each other and passes the buck as a reflex action. Or the second cart before the horse foible of constantly chasing new technologies and then the problem to solve it with...

Centralize versus Federated control of IT for complex multi-divisional, multi-country and subsidiary of companies?

• Right sizing IT solution for subsidiaries that operate at different cost base versus bludgeoning them with a corporate determined standard that's way too expensive for the low margin division, or countries with very low GDP per capita
• Allowing subsidiaries to write their own destiny in IT decision making but maintaining quality, risk management standards and procurement transparency
• Does leadership have self-introspection ability to adopt Enterprise Architecture? Where revenue and cost efficiency can be traced to business capabilities and processes that are intimately driven and powered by technology.

In summary, Milestone #1 can be done fairly quickly with fat easily identifiable and removed. Milestone #2 requires work, and that means investments for manpower to perform the necessary migrations and weight the immediate returns and long term cash flow as a result of migration.

Milestone #3 unfortunately, requires external help as hard questions result in even harder answers to swallow. It can be hit and miss as difficult to swallow answers can be watered down by consultants as they walk the delicate line of maintaining ongoing relationships with the customer...

Disclaimer: Nothing to do with MetLife. Had an interesting lunch with an old friend that needed some advise on the topic. So decided to brain dump how I'd approach it.