On Your Mark. Get Set. Go!

Before you even think of trying Compiere. What you should worry about. What are the issues? These are stuff that our prospects and clients went through. No holds barred. We tell it like it is.

On Your Mark. Get Set. Go!

Postby In Deep » Sun Oct 23, 2005 2:14 pm

WAIT!!! How many Compiere sales pitches have you heard about how you'll be up and running in 6 weeks? OK, maybe 3 months, at most? Well, let's see, I began writing up my specs in January, and what's today's date? Hopefully, we'll be in production by this January. So what's the deal? Do my programmers/integrators suck? Definately not. They're actually quite good. I'll tell you what the deal is: If you are a new business and/or have never run a ERP/order entry/inventory/etc. system, then, yes, you'll be up in days. But I tell you what, the company I work for has been in business for 60+ years and has been on their legacy system for close to 20 years. They are very set in their ways and are unable (or unwilling) to change the way they run the business. And you know what, they're changing a lot in order to run Compiere (mostly for the better, but not all - Compiere isn't perfect).

So why is it so different to run Compiere at a long established business (with a legacy ERP system) vs. a new company? Simply put, a new company will typically set up its business processes around its business application. After all, a young company doesn't really have any formal processes anyway. This is a point that I had to explain many, many times to the people I work with. They say, "It wasn't this hard 20 years ago, and surely technology has gotten better/cheaper since then." Yes it has, but 20 years ago, they (my company) changed their processes to fit their 1st system. This was easy to do after all; the business was smaller and the world was a simpler place. If something in the system didn't work just right, no big deal, because it was done by hand before anyway, so it wouldn't be much more work to do it by hand again. This is no longer the case. Obviously.

OK, so we needed to modify Compiere to fit our processes. No big deal, right? Well, that's right if the people that know the processes are clear and complete about how the legacy system works. This is very hard to do because it requires an in depth knowledge of both the legacy system and compiere so you can compare the gaps. Yeah, this get's a little sticky, but you know what, that's a whole topic on it's own, so let's table that for now.

Next, we needed to migrate all of our data. Easy you say because Compiere has all these built in import functions. If that's what you say, you are most likely WRONG. The import functions are pretty basic and if you are importing relational data or maybe just adding a field or 2, you need to make some changes to the code. Yes, more work. Maybe, you say, why migrate all this data. Start fresh. I say good luck. I personally think this is a lot of extra work. But you know what, maybe I'm just being judgement. Just because I didn't do this, doesn't mean it's wrong (yes it does, I'm just being nice Wink. If you can convince your company to do his, congratulations, you've done a much better job than me.

So you're probably asking yourself, why did I put myself through all this pain? Why don't I just give up? Well, because what you should probably know is that all ERP migrations are going to be this much work. Talk to other people who have done it. The people I know that have done similar projects all say the same thing (regardless of the system): THIS IS NOT EASY.

So I leave you with these thoughts:
1-The majority of IT projects fail.
2-There is no such thing as too much planning.
3-Get some good help (and a big budget).
4-You do not define the requirements; Your users do. And maybe your boss Smile (and to answer Enrique's question, my users required that all the features that existed on the legacy system be functional on Day 1 on the new system (Compiere). New functionality (vs. legacy system) could be phased in whenever I want, which will probably be after all the users are comfortable with the system.)

Anyway, I'm sure I brought up a lot of "controversial" subjects. So respond and give me your feedback. What did you do? Remember, if you don't say anything, I'm just going to keep talking.

See ya soon.
In Deep
Posts: 3
Joined: Sat Oct 08, 2005 6:16 am

Postby Guest » Wed Oct 26, 2005 5:47 am

Hi In Deep,

I agree with most (if not all) of what you said. I’ve been through a few large ERP/Accounting migration projects with government and higher education systems and it is a lot of work at best and virtually impossible at worst. However, I also think that migrations are usually much more difficult than they have to be for two reasons.

1) Users refuse to change their work habit or procedures to fit the new systems work flow.

2) Users want their old data migrated to the new system verbatim rather than summarized.

Obviously, some data must be migrated verbatim (products, current customers, orders etc) but some data, mainly historical data, can either remain on the legacy system until no longer needed or better yet, summarized and loaded into the new system. And some data may simply not belong on the new system because it is no longer relevant in its current form or current business (I hate to use a buzz word) paradigm.

I personally believe that a lot of migration projects fail or go over budget for one or both of the reasons above. I’ve seen very useful software become virtually useless within a few years because site specific changes were so numerous (and undocumented) that upgrades became virtually impossible and functionality quickly degraded as new bugs appeared after each upgrade.

I used to be a Unix systems admin/Oracle DBA and decided to go the “Entrepreneur” (read as self unemployed) route about 4 years ago. Things have been good, more work and risk than being employed, but rewarding.

So now, after 4 years, I am implementing Compiere for our company. There is only 4 of us (2 employees and two owners) and I am the only tech. We do E-commerce exclusively. I’ve been looking at Compiere for a few weeks now and it is just now starting to make sense. It’s really humorous, I used to work around ERP systems for years and in 4 years I have just about forgotten everything I knew.

We used to schedule 12-24 months for large migrations. In my opinion, 3 weeks would only be possible with new installations, and even that is optimistic. Not because of the technology factor, but because of the human factor. People need time to absorb completely new business work flows and concepts. I’ve estimated 6-12 months for me to implement Compiere in our company. I am hoping that I can get it done in 3 months, but I learned long ago, the hard and painful way, make your time estimates based on worst case scenario and hope for the best case scenario. :)

But anyway, Compiere seems to be maturing quite nicely. I’ve looked at a lot of the offerings out there and Compiere seems to be the leader of the pack for our needs. Like a lot of people, the thing that kept me from looking at Compiere earlier was the dependence on Oracle. Don’t get me wrong, I think Oracle is one of, if not the, best database systems out there. However, it is over kill for some needs and quite expensive. I’m not sure what the Oracle license costs are these days, but when I was in the industry I was routinely astonished at how much it costs to license and oracle installation. Of course, Oracle now owns InnoDB so who knows what is going to happen. :) But I like Mysql also.

One of the things I really like about Compiere is its modular design (OO perhaps). We currently use Interchange [http://www.icdevgroup.org/] as our shopping cart software and it is designed similar to Compiere. (very modularized) so site specific changes can be completely separate from the baseline code. Upgrades are then VERY easy. So I am hoping to get Compiere installed and be able to apply new upgrades with a minimal of effort.

I have not used Compiere in production yet, and my experience with it is only analytical, but so far, I like what I see and I’m willing to invest some serious time to give it a try. I just hope I can sharpen my skills enough to be able to give some code back to the Compiere project. (I’m a PERL guy, Java is new to me)

Postby tender » Mon Oct 31, 2005 1:19 pm

i am doing migration now to comppiere for bout a month, and whole-heartedly agree with EVERYTHING In Deep said. man this is not easy for a legasy system. it need a lot of hard coding. the data dictionary helps sometimes, but more often than not become an obstacle to flexibility. we are working on 'young' legacy systems with limited functions in the first place, and it has not been easy. i respect you guys who can tacke big, old legacy systems with a lot of functions. sometimes it feels that customising compiere is harder than building a system from scratch. i know the CORRECT way is to ask users to adapt to some of compiere's uber precess and workflow....but if we are that good at persuation we would get rich being salesmen. my schedule is one month. seeing all your forecast, i am not optimistic.
learning curve is high so far.
Posts: 5
Joined: Wed Oct 26, 2005 1:22 pm

Return to PreSales Bites

Who is online

Users browsing this forum: No registered users and 1 guest