Pricing an ETL tool 4

[Elsebet]

Currently I am required to make the case for bringing in an ETL tool to use in our Oracle data warehouse. All of our code is hand-written in PL/SQL and we have no metadata management currently, so I have plenty of ammunition technically to justify the tool.

However, the problem I am having is financial. I've spoken with and e-mailed plenty of vendors, detailing where I am in the process (no budget or project yet, just feasibility) and asking them for a ballpark cost figure for their tool at the low and high end. I'd like to put into my proposal a general cost figure associated with the more robust high-end tools like Datastage/Informatica down to the cheap/free tools like DTS/OWB.

However every single vendor (Oracle and MS excepted, costs are right on their website) dodges the price question, instead insisting they come in for a demo. I try and explain again where we are in this process and I usually get no reply afterward. Are ETL tool prices some sort of secret? To me this seems like poor customer service, after all I am trying to make a case to buy an ETL tool from one of these vendors who refuse to even give me a vague idea how much it will cost.

I did plenty of google and gartner work but the only figures I can find are $200,000+ for Informatica and $180,000+ for Datastage. What that amount gets me I'm not entirely sure. OWB is $5k and DTS free since we have SQL Server in-house. I'm so far failing to find an idea of how much the Cognos, BO, and SAS ETL server cost and like the "big guys" they are dodging my price questions. Anyone have any advice?

 

[MDXer]

You will have an extremely difficult time getting prices for these types of products until you have Either commited to purchase one or can give them a deployment timetable. Many of these companies want to do the demo because they feel they can make a better case for the cost by demonstrating the power of their tools to those who will make the final decision.

The best way to get the information you are after is to start a Formal RFI (Request For Information). Take the top dozen apps your looking at and develop a detialed list of needs. The create a formal document stating the Intent of the RFI the terms of a vendors participation and the time frame that response for consideration will be accepted.. Then break it into sections and pose the questions in a manner that will get the answer to a detail you are looking for.

Such as:
1) Security
1.1) Does the application integrate with current network security or does it maintain internal security.

1.2) if integrates with network security what types of security models does it integrate with. (example Active Directory)

2) Database Platforms

2.1) What platforms can the application work with

2.2) Does the application maintain full Features against all Platforms

3) System Requirements

and so forth.

Contact the vendors and give them a heads up to your process and findout who the area rep is so you can make sure it gets delivered to the proper person. Feel free to include requests for web demos and the possiblity of an onsite demo at the vendors expense. Once you have norrowed the field down using the RFI the you can move to the RFP Request for proposal in which you request pricing and possibly discounting models such as Higher Maintance lower Purchase. Again make these as requests the will gie the details desired.

I headed one of these processes for a BI Reporting and AD-Hoc query tools. We weren't the largest compny out there and still had vendors willingly bending over backwards to get us the information we were after.

If the vendors ask who they are being compared against feel free to let them know at the RFP level. Often a vendor will price their product lower than the cometition to secure the sale. Cognos once told me they would garuntee the would not lose our business based only on product price.

This type of process is usually more effective than casual enquiries because it lets the vendor know that their is an identified need in your product as-well-as that your company has a serious commitment to idenifing the right tool for it's needs.

Shoot Me! Shoot Me NOW!!!
- Daffy Duck

 

[Elsebet]

Thank you MDXer, that helps greatly.

I was sending out a high-level plan with my e-mail inquiries, however it was nothing on the level of an RFI or RFP. I do have a requirements matrix set up and a list of deal breakers which should help in creating the RFI/RFP. A google search already gave me a few helpful templates for both documents.

 

[MDXer]

Don't skimp on the amount or level of detail of information you are requesting. My RFI was 32 pages and completed documents were between 50 and 60 pages. A number of vendors commened that it was the most detailed request they had seen. which ultimately aided them in providing a greater detail of information and predicting what to do as far as demos. Don't be afraid to make it on their dime either. A sale at this level is worth a few dollars on their part.

Shoot Me! Shoot Me NOW!!!
- Daffy Duck

 

[mm0]

You can get the basic pricing information for list pricing from the vendors. Online you will probably not find it, and I don't post those slides from my presentations because they go out of date fairly quickly. Here's a rundown on the top vendors as far as features/functions:

Basic versions (departmental/stripped)
Ascential 90K, ~20% for annual support
Informatica 90K-100K, ~20% for support
Business Objects (Data Integrator) 40K

Enterprise /full-on versions including all major features:
ASCL ~425K, 20%, includes metadata, profiling, DQ, only DataStage engine ~225K
INFA ~350K, 20%, includes metadata, profiling, DQ, only PowerCenter engine ~200K
BOB ~275K, 20%, but DQ and full metadata repo via partners, only Data Integrator engine 150K

Oddball case:
SAS, which isn't close in usability, manageability, feature set (very much like a 1999 version of one of the others), costs 2-4X as much depending on what modules you buy / already have if you are a SAS customer. Big thing is to remember SAS charges ~45% license renewal fees which means your 3 year cost is almost always higher than any other vendor.

 

[mm0]

One additional note on the prices I listed. These are based on list price. Assume discounts if you want to know what you will really spend. The vendors will also mix and match components if there's something important to you - this is a very competetive market. Generally you can expect anywhere from 15% to 30% discounting, depending on what you are buying, what other vendors you are looking at, etc.

 

[timster11]

Great info guys.

I work within a small data department & have been asked to review ETL tools/software. We currently use Visual Foxpro to import, format, cleanse, merge & export our client's data. Some clients provide us with flat file data extracts - other times we extract the data from RDB we maintain on behalf of clients.

My objective is to review the benefits of introducing ETL software to our department (and therefore replacing much of the current processes & procedures we currently complete in Foxpro_. I have a ball park budget of up to 10K. I have not heard of these tools until today - so I don't know much about them.

Any suggestions on possible low cost ETL software solutions would be very much appreciated!

Thanx

 

[MDXer]

timster11

The best case you can make is ROI if your budget is flexible you may be able to make the arguement that a larger investment in technology can save Thousands in lost time.

10K investment saves 1K in lost productivity might be a hard arguement to make

20K investment saves 60K in lost productivity a manager would be nuts not to lobby for an increase in budget.

typically the more expensive the more robust it is (not always but often) so you need to look at what your ETL tool needs to support and how much effort you will need to replicate.

Shoot Me! Shoot Me NOW!!!
- Daffy Duck

 

[JoinerX]

Timster,
I agree with MDXer but would also add the following.

Whoever's paying the bills has to consider (among other things) the following:-

After I cover all the bills from now until next budget, including cockup allowances, how much have I got left and what is the priority list of purchases that fit in that budget.
After that, they will want to compare the expected labour cost of new client tasks over any purchase. Even if it would save money, if the difference is not substantial enough to offset any learning curve/implementation cost it may not be considered.
After saying all that, money is not everything. If you can show that an ETL product will improve customer service and free up resources but take 2 years to pay for itself, it may still be a viable option.


You may find it helpful to do the this...
a) calculate the approximate cost of labour per year that is spent on the small tasks (i.e. the salary X the hours) then double the figure you got and that may approximate the real costs per month. You now have a budget to work from. (so if it’s $20k you won’t be calling Ascential for a demo
b) Build a prioritized list in descending order of what takes the most time for each job. Then build a list of which client tasks take the most time. It may be that some clients are far more valuable and represent a large chunk of the turnover but their ‘tasks’ do not take much effort. If there is no conflict there… continue
c) From your prioritized list of what takes the most effort, create a functionality requirement list for any product you buy. Email that list to the sales department of a few vendors and let them do the running around.

It sounds like anything over $20k is out of the window. In which case what you may need could be anything from a script generator like Wherescape, to basic ETL/data feed tools like Scionet DTS, Pervasive Data Integrator and I think the bottom end of DataMirror might be your top end.

As MDXer pointed out, the time and cost of supporting a product of this nature can out way any initial benefits you may gain. To be honest, I’d be looking at pushing some kind of standard (if possible) onto the clients or at least planning for one.

Good luck.

 

[yksvaan]

I agree to everything above. Here are few things to consider too, thoughts more than tools. Maybe these are obvious and you've thought of them before, but I like to point them out and I think they apply even if project would be of a smaller size.

I am not favoring some of the mentioned products and I'm familiar with only some of them.

Cost of labour (overall, including training, recruitement costs, maybe consultants, etc) : the hours and the level of cost per hour varies depending which ETL-product you choose, which make evaluation more difficult. These basicly depend on two factors, how much training is needed to get to a reasonable competence, and how many professionals are available out there. I have no formula for that, but if you have the technically greatest product you can have per dollar, it's no use if you can't occupy the seats in the project. Here comes in of course that what kind of competences you have in your firm already. As you can see looking at different forums on tek-tips, mastering any product, even on some problems that look trivial at first, can take effort.

As we all know, in projects specs/definitions from customer/application user-organization change during projects and even in the final stages of some case, you may have to go back almost to the beginning. How well does etl-product cope with this? Can you get the first test versions out fast, so you get the customer to tell the obvious changes? So what kind of development cycle is typical using certain product? I think this is a major financial issue too.

How fast you can have the first reports/end-user applications out? The sooner the likely your financer/ sponsor is going to budget money for this project properly. This issue may have a huge impact on how long the initial project is going to take, do you have proper resources instead of creepy budget which slow downs development cycle. So it's a financial issue. Let's be positive for a second, probably you won't get into a vicious circle.

What I am aiming at is that after few years considering the end-user part too, the price of the product is not the major cost in overall cost, so one shouldn't make a choice just comparing prices. I don't mean you are doing that here.

One more thing, the vendor is probably thinking, what could they sell you after this. You could use this as "leverage" when negotiating with vendors.

Cheers