Outsourcing II–The "What" Question

A little while ago, I was asked to give a presentation to CEOs on outsourcing. The audience wanted to know about adopting outsourcing for their companies; making use of its promise while avoiding its pitfalls. It seemed to me (unimaginatively, I must admit) that the whole thing boiled down to four fundamental questions - the why, the what, the who and the how.

I decided to expand the presentation into a series of blog posts, one per question.

The What Question

The second in the series deals with the what – choosing which parts of IT can be and should be outsourced to a partner. Of course, one must first decide how one defines "parts of IT" in the first place – and different companies have slightly different approaches. Some will do it by business unit, others by geography, reporting structures or cost head. I have found it most convenient to consider any application, function or service as a candidate for outsourcing.

• Applications are software-hardware combinations, either purchased or home-grown, that assist business processes – such as ERP or Payroll applications. Most companies have these well defined.
  
• Functions are support activities carried out by IT not specific to a business process, such as desktop support, server maintenance or data security.
  
• Services are activities carried out by people in the IT department to support a business function. Common services are activities such ad-hoc reporting, rights management or  application development. The line between business and IT is often blurred in this set.

In many companies, an individual application or service will span cost heads or business units, and most times it is sensible to outsource the entire services (such as desktop support) rather than outsourcing it by business unit. In other cases, piecemeal outsourcing may be better – for example, national boundaries often prevent multinational companies from choosing a single payroll partner. However, the three classes above are better considering the what question; a grouping such as geographical or hierarchical is more appropriate in deciding who.

Instead of filling my lungs and spouting the mouthful "application, function and service" whenever I talk of what can be outsourced, I will arbitrarily shorten it to "service"; unless otherwise mentioned, this will represent all three and save seven syllables in the process.

Now that we have our savings done, terminology sharpened and knives ready lets get back to the business of carving up the pieces and figuring out which ones are worth considering. This consideration is, I propose, driven primarily by two factors:

• Is it possible to outsource the service (relatively) painlessly?
  
• Is it desirable to outsource the service. Are there noticeable benefits?

The answers to these questions are, obviously, not simple yes/no. There is a certain amount of complexity in deciding the boundaries of each piece, and the questions themselves have nuanced answers. A little bit of framework may be in order.

I've chosen two variables as the axes – Investment and Utility. Investment  here is defined as the amount of resources a company invests into a the service in question, while Utility captures the value this service delivers to the company. They are reasonably independent; investment does not guarantee utility, and utility does not always need investment – which gives us four proper MBA-style boxes to work with. Investment is also a proxy for potential savings from outsourcing – the more the investment, the more the quantum of saving expected. Utility acts as a proxy for risk – the more use an organization gains from a service, the higher the expected loss if outsourcing does not go as well as expected. The reason for these proxies are that actual savings and risk are nearly impossible to estimate directly.

Some people will recognize Bruce Henderson's BCG Growth-Share Matrix in this, and indeed some of the terminology is derived from there. Be warned, however, that the BCG matrix had a very different purpose and was analysing a different subject using different parameters. The names may be similar, but the analysis  is quite different; this is not the BCG Matrix. I re-used the box names (except Question Mark, which I substituted with Work Horse), because most of them sounded right for my purpose; nothing else is common.

• Star – these are services that consume a lot of resources, but are at the same time very useful to the company. Usually these are the core revenue generating services, such as the trading engine of a brokerage, or the billing system of a telecom company.
  
• Work Horse – these services consume resources, but produce work that, though vital, does add much utility to the company. Typically these are necessary services such as security, infrastructure maintenance, vendor management which do not contribute to the bottom line but are necessary to run the company.
  
•  Cash Cow – these are services that have substantial utility for the company but consume little in resources. Many companies have services that they built in the past with little investment that turned out to be very useful. They neither require maintenance contracts, nor are they enhanced or re-developed regularly.  IIFL, for instance, had a system for chatting with customers that wanted to use yahoo chat. As a broker, it was a compliance requirement to record every chat, and someone had installed a simple open-source application that did just this. It needed little attention and in the three years that it had been in operation had generated crores of revenue through the chat channel.
  
• Dog – these services contribute little but don't consume much either. They are often niche services of legacy leftovers. Since replacing or substituting them is often high and utility low, these services are generally left alone till they die a natural death.

Utility is better measured as contributing directly to the revenues of a company, while Investment is similarly that which contributes directly to the costs. Investment thus is not just money – resources, management time, regulatory headaches, many other factors feed in. This framework is meant to be simple. "Utility" and "Investment" are deliberately inexact. If I were to replace "utility" with "ROI", then using this model would require an army of accountants. Companies often have a poor idea of the exact quantum of utility of or investment in a service. Faced with two services, companies find relative ranking (this is more useful than that) easy but relative rating (this is x% more useful than that) very difficult to answer.

When I was doing these outsourcing evaluation studies, I realised it was enough ask people to rank (without insisting that the rank be unique) and assume that a ranking is a rating (that the values are equally distributed along the scale). This is because this model depends entirely on which box a service is in, not where in the box it is. This approach also makes it different from the scatterplot of a traditional four-box analysis.

Once the ranking is available, most CIOs find it easy to say which service (and thus which rank) should be in the middle – the rest obediently fall into place. This approach is not very useful if you are a consultant; I did these studies for service portfolio for a division each of Reuters and Schwab - each took barely two person-weeks of billed time. For a CIO, however, this is quick, dirty and very effective.

Here's how the boxes map to identifying outsource candidates

Type	Outsource Strategy
Star	These services are usually the glamour investments in each company, and they are often implemented by partners. ERP, CRM, SCM, all these are start candidates – large IT expensive projects that when properly implemented lead to substantial gains in revenue or operational efficiency. Because star investments are large and specialised, they usually require a large injection of skills in the short term that only a partner can provide. When properly done, however, most star investments should be part of a larger strategic thrust. They are often not part of such a larger plan, and thus often fail to deliver on their promise by consuming a lot and not bringing as much utility as expected.
Work Horse	These are ideal candidates for outsourcing. If considering the tactical approach, the primary benefit here is cost and predictability of service levels. As part of a strategic approach, the intention is often to get rid of these activities altogether, transferring the responsibility itself onto the outsource partner. Take CRM, for instance (usually considered a Star). A CRM implementation has the associated workhorses of server maintenance, license management, upgrades – these are easily and usefully outsourced to a tactical outsource partner. However, a strategic call may be taken to move to something like Salesforce.com – in which case the associated work horses are removed entirely.
Cash Cow	These are usually best left alone – do not fool around with what is working well. While they may be good candidates from an operational point of view, the risk of change – things going wrong at transition - is high and the rewards low since investments were low to start with. A company with a lot of cash cows is in a good position – tactical outsourcing is not relevant and one can focus on analysing the strategic. However, cash cows do not remain cash cows for ever – often deteriorating into dogs as their utility declines.
Dog	These services are of little interest in either the strategic outsourcing scenario or the tactical one; they are usually best ignored. The one caveat here is if a company seems to have a lot of dogs. Remember that the matrix is the company's perception of the classification of applications – if a company perceives that lots of services are consuming little but delivering little, then possibly something bigger is wrong. Possibly a company is under-investing in IT, or is excessively pessimistic about the application of IT.

Deciding this is not entirely divorced from the why and the decision is usually not sequential. Once can arrive at the why after having chosen what to outsource, or sometimes the decision on why can drive the choices available for the what. One key way in which the what question is tied to the why is in ensuring that choices match intended direction. If you are planning a strategic outsource, you will be looking at outsourcing as large a pool of services as you can. If many services are not suitable for outsourcing, then you will not gain as much from strategic as you are planning. If, on the other hand, you choose tactical outsourcing but almost nothing is in the work horse quadrant, you might want to look again.