It would be nice to think that ECM always comes with a positive and sexy Return On Investment (ROI) making everybody involved with the purchase and implementation of an ECM solution happy campers. However, if ECM had such a positive ROI then ROI would be one of the lead reasons for purchasing an ECM solution but that is rarely the case. Is that because ECM doesn't have a positive ROI to talk about?

No, that wouldn't be a fair statement, some of the applications made up with ECM are very capable of providing an easily calculated ROI but those are NOT the applications usually pushed by ECM vendors. The reason is that about the only way we know how to accurately calculate a ROI is where there are hard costs involved ... Archiving of any kind often provides a positive ROI because it relieves more expensive applications of their resource requirements.

Email archiving tends to significantly lessen the burden that email servers sometimes feel under load .... this is really only true for costly groupware servers like Exchange and Domino because they are themselves databases that have scaling problems. An archiving service can prevent the need for a second (third or fourth) email server and the hardware, license and administration costs associated with it.

File system archiving can relieve the pressure on expensive Tier 1 storage and utilize the much cheaper Tier 2 and 3 storage devices, there can be two full orders of magnitude difference in costs between Tier 1 and Tier 3 storage and the more content you can move to cheaper storage the cheaper storage becomes (though to be honest, there are better ways to achieve that ROI today than any ECM vendor can provide).

Application archiving (like SAP) can relieve the pressure on the database that the application rely upon, this usually means fewer servers and smaller databases which means less hardware, license and administration costs associated with bigger systems. It isn't really the ECM archiving application that is supplying the savings, it is the relief of much more expensive business applications.

All of these archiving solutions provide an ROI because the underlying system they are archiving from are more expensive then the archiving engine. The problem for ECM vendors is that it is that it is difficult to make the expected/hoped for 30+% profit margin on archiving solutions because of the ROI calculation ... if the ECM solution is too expensive then the desired savings are simply not going to be found. So the costs of archiving software stay relatively low but then the solutions themselves staysrelatively static.

Contrast this with the business side of ECM, workflows & BPM, Email Management & Monitoring, searching and Knowledge Management, Records Management and governance .... all of these solutions CAN lead to improved business processes, greater productivity, and stronger governance. In some cases these improvements can provide savings that far exceed the costs of the ECM application  and indeed anything that a simple archiving tool can bring, but you will rarely if ever hear an ECM vendor touting ROI for these applications because the costs of the ECM solution are themselves very significant and relatively easy to determine but the costs involved in the existing processes they are replacing or improving are usually not well understood at all and likely underestimated.

Also, whereas archiving provides a hard savings with a formula and a way to check results, most of the other ECM applications provide soft savings. The hoped for savings come in improved productivity or "less errors" or some other preventative measure like lowered legal fees. Examples abound:

• A business process that once took 3 hours of effort now takes 1 hour, the expectation is that the two saved hours can be calculated as productivity increases ... but there is no way to enforce the use of those hours nor any easy way of showing that there are in fact 2 hours to be found (it could be that those hours are just shifted to somewhere else within a larger business process) so the savings are usually intangible and may not be realized. 
• A business process that once saw a 5% error rate could see the errors reduced to <1% providing real, tangible benefits; though it is possible that new errors can be introduced because of the new technology and not recognized as such ... for instance a form that was once contained a free-form comment now has a restricted textfield and that field had routinely but rarely been used for some undocumented purpose, or more realistically that the errors that are in the new process are magnified many times over what they would have been with the old process (computers are fast!)
• A properly conducted and maintained Records Management program is one of the most expensive corporate programs whether or not the RM is managed and enforced by an ECM solution. Nonetheless, a properly functioning RM program can serve to significantly reduce the costs associated with the legal discovery process, work that if left unchecked can double, treble or more the external legal fees associated with a lawsuit of any kind. Lawyers tend to work at rates exceeding $150/hr, often well over $400/hr, and RM can reduce those charges to a small fraction of what they could be.
  
  • Of course you have to get sued first with the kind of allegation that leads to wide-ranging discovery orders and scads of content, those industries that tend to get sued a lot (say big tobacco) have little difficulty in finding the ROI but less controversial industries that don't tend to get sued much take a lot longer to find that ROI.
• Search is supposed to help you find things faster and lead to more knowledge spreading (and more knowledge spreading is supposed to translate into wiser business decisions, though there has been no study that actually shows this link, it is just assumed to be so) but with more and more content being searchable it takes longer to determine what is relevant and with search technology being so cumbersome to administer, it is far easier to just index everything and make it available to search ... but that means less discrimination from end users and may lead to poorer and longer decision making.
  • It is very possible that the "ranking" of results drives thinking processes in a way that wouldn't happen otherwise, there is no way of knowing whether that is positive or negative and no reason to believe that it would be one or the other -- blind faith.

As I say, there is always the possibility of a positive ROI and there is the dream of a very large and very positive ROI for many ECM processes, but the reality is that there are damn few case studies available to back up these ROI claims. Even those corporations that are sure there would be and are convinced there is a ROI do not do the work to determine what it turned out to be and probably wouldn't really be able to accurately do so in any case.

A good independent ECM trusted expert should be able to look at the overall business and determine whether or not a measurable and significant Return On Investment is likely for any particular ECM solution deployment. It is very much on a case-by-case or at least an industry solution-by-industry solution basis.