This is the second in a series of three articles intended to assist with cost factors and return on investment (ROI) for e-discovery. Read the first installment here. The third article will deal with the metrics and cost of review.

E-discovery can be daunting and costly. We look to reduce review costs with predictive coding and advanced analytics and we look to reduce technology costs for processing, hosting and storage. Bringing e-discovery in-house is the right move for some but may not be for others. Sometimes, a hybrid solution works best. Since some acquisition costs are not very obvious, here is a model to help you make your decisions:

 1. What’s the current cost?

What do you currently spend on e-discovery technology and what are you looking to save? Look at historical spending from past vendor and law firm invoices. For the latter you may need to tease the technology costs from the legal costs. Compare these with your projected caseload going forward to determine your average annual spend.

 2. What are the priorities? 

Do you want to cover the whole Electric Discovery Reference Model (EDRM) or just litigation hold, preservation and collection? Most organizations start “left” in the EDRM and move right as budget, experience and caseload increase. Do you plan to implement review capability in-house, say, for HR cases and to outsource the rest? Before you start putting out requests for proposals (RFP) and looking at vendors and software, make strategic decisions on what you want to build, how the pieces fit together and how you want to grow e-discovery in-house.

3. Is there a strategic plan for IT?

Forget for a minute that you need support from the procurement or finance departments. Is there an IT plan with components you can use? Is there movement toward single-instance email archiving, document management or enterprise search? You might find savings by combining effort.

4. Who are your users? 

If your large department has, say, divisions for commercial, IP and HR litigation, are you planning for each, or are they operating independently? Do you expect to have outside counsel log into your systems? Whose buy-in do you need, and if you don’t get it, will it affect your ROI? If you are in a smaller enterprise, do you have the will to embrace an in-house system? Answering these questions help to determine the size of the e-discovery enterprise you will build.

5. Who creates and defends the process? 

Too often, businesses answer this question too late. E-discovery is as much about lawyering as it is technology. If you are going to preserve, collect, store, process, cull, review or produce, counsel must explain and document how the organization will meet e-discovery legal obligations. Someone must be trained and available to testify. Paralegals, vendor and support staff can help design and optimize a workflow, but counsel must be accountable for the discovery response, both institutionally and for each case.

6. What about software, hardware, training and staff?

These are the hard and obvious costs. The pro forma illustrates license, license term, maintenance, updates, servers, storage, installation, outside staff training, inside user training, ongoing training, add-ons for advanced tools such as analytics, and minimum levels of staffing to sustain an in-house program. Of course numbers will vary depending on your priorities and size. If the numbers surprise you, recognize that any serious capability you may want in-house is much more complex than just desktop software. You cannot count on IT just to come over and run this without an IT service level agreement; you will need your own client-facing staff and you will need dedicated IT staff to run and maintain the systems. Below is a sample list of costs:


Software licensing, maintenance, support

$ 50,000

Per Year

Advanced tools licensing, maintenance, support

$ 25,000

Per Year

Servers and storage

$ 80,000


Process development and documentation



Installation charges

$ 20,000


Two e-discovery and one IT professional

$ 225,000

Per Year

Training charges for in-house professionals

$ 10,000


Training time and charges for usership

$ 30,000


2nd and 3rd year training charges

$ 15,000

Per Year


Year One

$ 440,000


Year Two

$ 315,000


Year Three

$ 315,000


Three-year costs not including internal implementation time

  $ 1,070,000


7. What is the time and cost to implement? 

Implementing e-discovery in-house can take months or years. How much time will you, your staff, IT or procurement spend to do it? You need time, money, research and energy to document the defensible internal process. You need the same to evaluate RFP responses and to select, test, prove, contract and implement with the right partners and software. These activities make all the pieces fit together, so consider their costs.

8. What’s the total cost of ownership (TCO)? 

If you do this homework you are ready to develop a TCO. Compare it to the current cost (see chart above). Over one to four years (because it will take a year to implement) does it recover the intended costs? If the argument is not clear financially, do you have other legal, strategic or tactical reasons to build e-discovery in-house?

9. What about longevity and flexibility? 

What if you spend a million dollars and more than a year in planning only to discover that the technology doesn’t keep up in two or three years? You do not want to find yourself with out-of-date software or methodology. How long should it take to return your investment? The solution needs to be flexible and scalable. Will you be nimble enough to easily incorporate more cases or new technologies? Get independent research and use more than one crystal ball.

10. Are you ready for the decision? 

Don’t be surprised if your calculation points to a hybrid solution.  Many of the busiest and most litigation-ready law departments build heavily onto the left end of the EDRM but outsource or contract functions for which software and infrastructure are the most costly, for which the demand fluctuates considerably and for which the obsolescence is unpredictable.