General discussion


ROI of adding memory

By brad_mankoff ·
A few years ago I did research on the cost of adding memory to computers versus down time or just slow computers. I would like to find a spreadsheet that I can use to justify increasing the memory in one of my customers computers and showing the powers to be that this little cost will save this much money in lost productivity.

Second is to justify purchasing new computers with 512MB RAM instead of the 256MB one of my customers currently do.

This conversation is currently closed to new comments.

Thread display: Collapse - | Expand +

All Comments

Collapse -

Dont upgrade.

by tbragsda In reply to ROI of adding memory

I know your looking for solid numbers, Im just offering my two-cents.

Depending on the numbers of WKS, makeup of support etc, the cost of upgradeing ANYTHING at a desktop is simpley too much.

You have to look at your WKS lifecycles planing etc, but order enough memory to last the cycle, and dont plan on upgrades. If its any number of wks, sending techs around to install memory will cost more than the extended life you get out of the WKS.

Im sure its differend for many, but in every place I have managed, it has never made sence.

Collapse -

I partially agree

by LiamE In reply to Dont upgrade.

I agree that its best to get the memory nailed at the planning stage but lets be honest, stuff can happen half way through a cycle.

A service pack to a crucial app may well cause any given ammount of memory to go from sufficient to insufficient.

As long as you have the man power to hand you can send the IT whipping boy round to stick in the memory at a rate of perhaps 50 machines a day. Negligable cost really. And his wages are being paid anyway.

Whether its cost effective comes down to 4 factors. Cost of upgrade, time to PC replacement, time saved per user per day and cost of the employing the user. Lets say the cost of the memory stick and 5 minutes labour per desk comes in at ?100 (its more like half that with in house labour at current prices) and the machines are 18 months from expected replacement. If the cost of employing the user is ?25 p/h we can analyse it thus. We need to save that user 4 (240 mins) hours or more over the next 18 months (roughly 2880 working hours, 72 working weeks and 360 working days ) for it to be a good idea.

Anyway when the maths is all said and done that works out to the need to save 40 seconds a day to make it cost effective if that time can then be spent working productively.

Not a lot is it? Frankly you can save that on boot up on most PC's by adding more RAM.

And on top of that you get the not so easy to quantify 'happy user' benefits.

Collapse -

Good post.

by tbragsda In reply to I partially agree

But I think your 50 wks/day is too high. Send a tech around to do anything, and he will hit delays like;
1) users asking for help on the spot
2) waiting for "busy" clients that will make the tech wait 10min while they finish....
3) Chating with users
4) Problem wks could add 15-20min

I also think the "add 40 sec/day to make cost effective" not exactley correct. That presumes people work straght through their days. Thats just not people work.

The best you could get out of the upgrade is improved user satafication. The frustration/preseption user problem is worth a lot, and im not denying that. That in its self may be the metric to make it worth the time/money.


Collapse -

But if you do it while doing other work

by jdclyde In reply to Good post.

then it isn't a special trip and almost no time added for the memory upgrade.

How many times do you have to go to someones system for one thing or the other?

I know many people have remote access to push updateds down, but does that completely keep you from getting out and doing a hands on job?

Also, I think that the users like to see techs around to ask them if they are having issues that they wouldn't take the time to report but would tell you to your face. Many are just complaining without reason, but some are very valid. You have no idea how many times I hear from people "I didn't want to bother you about XXXX". Hello, that is why I have a job, to be bothered by you if you have ANYTHING stopping you from doing your job.

Collapse -

Yes, it could be a good time to see whats

by tbragsda In reply to But if you do it while do ...

Yes, it could be a good time to see whats bugging the users. Nothing wrong with stoping by one-by-one to see whats up.

If I were doing this though, I would send the tech round off-hours.

Collapse -

I pretty much agree

by jdmercha In reply to Dont upgrade.

Adding memory for a performance increase is usually not worth the time.

On the other hand, if you can't run a required app without more memory, then it is worth it.

Collapse -

Can you demonstrate it on a couple of PCs?

by gralfus In reply to ROI of adding memory

If you can show that it really does improve performance on a few systems and certain applications, then you have real numbers to take to management to justify upgrading all of them. (Though certain workstations use memory that can be prohibitively expensive).

Collapse -

Here are a few things to concider - then upgrade

by jdclyde In reply to ROI of adding memory

What is the cost of memory for the systems? Slots provided it is usually cheaper to buy a bigger chip and then replace from one and combine in another.

Have you or will you be adding new software or upgrading soon?

Do you get many calls for people locking up or crashing?

Example, a few years back 64 MB was fine for systems doing word processing and e-mail. Then the update to Lotus Notes comes out and it wants 128 MB, and is a dog without it.

Do you buy all new computers? Heck no.

So we checked our database to see what we had for computers and made a list.

we ordered a box of 128 MB. Would go to the first system and take the 64 MB out and put the 128 MB in. The second system I would drop that 64 MB in.

Problem was solved with barely more time than a reboot. Note, I used this as an excuse to **** the systems out so I got double duty out of my time.

The other part of crashes. Most of the times Windows crashes is because it ran out of resources. The EASIEST way to fix that is to throw some memory at it.

And yes, everytime the user has to sit their and wait on a computer adds up faster than the cost of memory.

As long as they are PC133 or higher, do the upgrade. PC100 or lower isn't worth the time and are very expensive.

Related Discussions

Related Forums