Ask HN: Have you ever saved your employer big money through simple change?

Many times. A few of my favorites:<p>- We upgraded our hardware and our forecasting software vendor wanted a one time $600,000 charge. I convinced my boss to replace them with in-house written software. Took 6 weeks to write.<p>- Our 400 worker factory was $30,000 under-absorbed per month. I wrote both standard costing and data collection software. Supervisors compared the standards to the actuals to discover where they were losing money. We were over-absorbed by $30,000 per month 6 months later.<p>- We budgeted over $1 million for a new ERP system to "solve all of our problems". I helped others solve most of their problems by identifying them and coming up with solutions from the existing software. We never did buy new software.<p>- (My favorite). Our HCFA feed from the U.S. Government was broken and no one knew why. I dug in and changed 1 byte of code (1 byte, not 1 line). The next day, our bank account had $6.5 million more in it. I never had the heart to tell them how easy it was to fix.

I took the CTO position at a company AccessUSA now HotelBeds which is a travel consolidator, they had a very old system built on fox-pro for the reservation system. The original developer had pretty much abandoned any updates to the system and was just sucking licensing revenue out of it. It was a totally closed system, so it could not be modified from an external development team and all advances made towards the original developer met luke warm responses.<p>Several consultants had came in and told the CEO that there was no way to graduate out of the system, that it would require a total rewrite of the system. The CEO had went through 3 failed attempts to build a new system with external vendors the second of which folded shop and then the CEO got a call from an Indian company saying that they had been working on the system for over a year and had not goten paid. Apparently the front company that folded was subbing everything to offshore firms without letting the customer know.<p>Anyway, all attempts failed and the one most critical business issue was that the rates and allotments for hotel rooms needed to be available to customers like Hotels.com, Travelocity, and Expedia in real time. It was a read only problem. I set up an export from fox-pro to a database cluster (I can't ever remember which one now (Oracle or MSSQL). Built some web services on top of it to expose the data and published the WSDL. All tolled we (2 people full time) spent a week building the services and added over 30 million in revenue, the first year. I have done some others, but that is probably my best time to money ratio of my career.

I once worked in an outsourced tech support call center for a top 3 US ISP.<p>One day we started receiving a massive flood of calls all about the same problem: the default ISP homepage wouldn't load, and people thought their internet was broken. The call queue was growing exponentially, and no one could figure out what was wrong. This went on for a few hours, with no solution in sight.<p>Long story short, while troubleshooting with one of these customers, I found the source of the problem (a new firewall setting in the ISP's free av/fw package), reported it to my supervisor, and a fix was quickly implemented.<p>Given the volume of calls and the size of their subscriber base, I probably saved the ISP at least 5 digits in phone fees, and got a lot of angry customers off their back.<p>For my efforts, I was rewarded with a company branded pen.

I saved a client tens of thousands of dollars per year by <i>not</i> making a small change.<p>The client had just had a customer service issue where one of their customers discovered many months later that her card was still being charged for her subscription even though the card had expired. She didn't know that many cards continue to work after they expire, and had never bothered to cancel her account.<p>So the client wanted me to make sure that we don't charge expired cards in the future. As far as I knew, this was the first time the issue had come up in the more than a year since we switched payment gateways, so I generated some new reports, and it turns out they had charged almost $4,000 to expired cards just in the previous month. Some of those customers would correct their card data if we notified them, but the reality of subscription businesses is that there's a lot of inertia involved. It's safe to say that the client would simply lose a fair portion of those dollars every month if we stopped charging expired cards with active subscriptions. Much better to refund the one customer/year who complains that they thought an expired card meant they unsubscribed than to forgo all that revenue. Of course they agreed and we didn't make the change.<p>FTR, this is the same client that a few months ago was complaining that I was the only programmer they'd ever worked with who bills them for email and phone conversations. I'll be reminding them of this case if it ever comes up again.

I can't tell you specifics yet since they haven't published (expecting it sometime in November), but there is a client who will be something like six figures richer twelve months from now on the strength of two content edits to a ten year old blog post by their CEO. ("What do you <i>mean</i>, the blog post outranks the product for our most valuable keyword!?!?!?")

When I joined my company the app I took over development for was hosted on Amazon EC2 at about $700/month (Windows 2008). The app wasn't even done yet and was being used by about 3 people, so I switched us over to shared hosting at about $60/month. The shared hosting is even providing us better throughput and the app is now noticably faster than when it was on Amazon; and I don't have to worry about the details of securing/maintaining a Windows server box. Not gigantic savings, but still $7600/year for basically no work at all isn't too bad.

The internal level editor at a game developer I used to work for had a habit of crashing. Lots. Something like 10-20x per day, per user. Starting it back up and loading the level took more than 2 minutes, ignoring any lost unsaved progress. Saving took some time, too, and risked crashing. The level editor probably had around 30-40 users. (level designers and environment artists) I'll leave working out the amount of lost time per day as an exercise to the reader.<p>For some reason, fixing this situation was not a terribly high priority for those in charge. Someone did create a crash reporter that emailed a stack dump plus any user comments to a central email address. The quota for that account was soon full.<p>I decided to spend an hour or two here or there on going through the crash dumps. Within a week or so, I'd fixed the bugs causing 99% of the crashes. After rolling that out, we were still getting tens of crashes a day. Fixing 90% of those was pretty easy, too. When I left the company, we were getting maybe 2 crashes a week.<p>As you can probably imagine, this made me rather popular among some of the artists and designers.<p>The editor was fairly ancient, they'd been using it for well over 5 years. I can't begin to imagine how much effort was lost in total.<p>Incidentally, that sequence of events also cured my fear of large, horrible code bases. This thing was written using a proprietary extension library to MFC plus some creative use of DirectX. I think the only code that can scare me now are life-or-death systems.

I worked in a construction/real estate company as IT support. The way they did sales was they had a large excel file in which they wrote their leads every day, and each day they had to send it to each of the other offices (over dialup, it was a few MB) and sync the changes other people had made with theirs.<p>I wrote a simple CRM web app in my spare time, which every salesperson in the company just <i>loved</i>. It enabled them to see each client's history with a single click.<p>I asked for a bonus, the manager waved me aside saying "sure, build it and I'll give you $100". I never saw an extra dime other than the $5/hr I was making (I probably earned them thousands in sales that would otherwise be lost, not to mention productivity). They still use the system to this day.

Not saved but made them more. As head of ecommerce for a travel company I suggested small ux tweaks to their checkout process and they got a 25% conversion uplift. Worth something ridiculous like £5m per year. Needless to say I still got a poxy bonus which is why I now freelance :-)

When I was a senior in High School I worked evenings at a call center fielding extended warranty repair calls for a major retail chain. There were many calls about older Macintosh computers that would not turn on. Some would give the chime, others would not, but they all stopped before video initialized. The standard procedure was to send out a replacement motherboard (refurb from a contracted parts warehouse). These replacements would often suffer the same failure within days of being installed, if not immediately. No one in the call center seemed to know the problem.<p>I knew a little trick from my days in the school Mac Lab. I would have the customer flip the power switch, wait ten seconds, then flip it off and back on again (not too quickly). Often I would have the customer stand up and spin around or some other humorous ruse before the second step. The Macintosh would boot up and work fine. I would order a new motherboard battery ($2.50) and the retailer would pay 75 bucks for a tech to go install it. Still much cheaper than what they were doing before.

Do a rough NPV and bring this up in your annual review if they try to nickel and dime you on the raise.

I once saved my old company about 300k(dollars) in fraud charges. They were a telecom company and were experiencing massive fraud on some of their toll free numbers from Brazil and the Philippines.<p>Using a web app I developed, I suggested removing these numbers from all public sites and advising customers to call the help desk for them. You'd be surprised how much pushing I had to do and how much resistance I faced, especially since I was working in the marketing dept.<p>Anyway, they finally implemented my idea. Total fraud for 2009 was approx. 300k. For 2010, its approx. 50k.

During summers in college, I worked for a small web dev company that had a private line into a client company's network (a microwave link between the buildings, encrypted and with firewalls on both ends.)<p>One day, we completely lost the ability to log in from our building, and therefore, to do billable work for that client (though they had space for 2-3 of our employees to work in one of their offices.) The other company insisted they hadn't changed anything on their end, so it must've been something I did. I did research, packet logging, analysis, and so on, and couldn't find anything. This went on for days. Eventually I walked over to their server room, and found the screens said "Windows 2000" where they had previously said "Windows NT". Apparently they hadn't changed anything <i>except for the operating system on the servers handling logins</i>, which was clearly unimportant and not worth mentioning. Another five minutes of research and I discovered Win2K server used an additional service (triggered from their end) to log in, which was being blocked by their firewall and therefore never appeared in my logs.<p>I opened one port on my end, their firewall guy opened one on their end, and a problem that was costing us half of our revenue was solved.

I saved a company I was contracting for 5k a month in wasted adwords expense in the first hour of looking at thier account. They were happy to pay my fee. :D

I worked for a bank long ago. Realtime authorization requests and responses went through an IBM Series/1 box--at the time a stack of computing power in a 6-foot cabinet. It locked up occasionally, choking off communications to/from our auths system. Back then computers lived in a secured super-cooled room and we programmers were forbidden to touch anything but the keyboard. SOP was to call in a tech to "fix" it, meanwhile Visa and MasterCard would "stand in" on our auths, at great expense and greatly increased fraud exposure. After a few late-night emergencies I realized that the tech couldn't even do diagnostics without rebooting the Series/1. And he never found a problem after rebooting. That was at least an hour of stand-in time down the drain, mainly spent waiting for the tech to get in there at 2:00 AM. So I developed the revolutionary technique of hitting the forbidden Blue Button myself instead of reporting an outage. Worked every time, and we almost never went into stand-in after that. Management thought a firmware upgrade had fixed the S/1, when it was just the ol' finger on the kill switch.

I'd imagine that most people who have worked for larger companies for reasonable amounts of time should have plenty of these stories - after all that's how good employees generate their value, through innovation. I've worked on projects in the past that improve manufacturing yield by small single digit percentage amounts. On product lines with revenue in the tens of millions, these few percent really add up.

I've no idea on numbers, it was never quantified to my knowledge (and I don't think it really could be). However....<p>My former organisation had a collection of MI reports in their core application to let them track any number of sales metrics. These were all decidedly hard-coded and a pain to change; manually added to the menu screen via source code and linked in with Crystal Reports templates. Frankly I don't think anyone in the IT team liked using Crystal, particularly as the reports were largely just straight results tables.<p>Myself and a colleague were less than fond of this system to put it mildly. I had new business area I was supporting and they needed reports. So, we talked to them about what they wanted and how they'd use them. Discovered Crystal was just being used to export to Excel, so....<p>One database table was set up to let us bang in new reports without code changes. One class was added to format results sets from the database into a relatively tidy file in Excel.<p>Total time to do those two changes was, I reckon, less than the total time to add a couple of reports the old way, and gave much faster deployment of new reports. On top of this, over the next few months this business area gained more MI reports than the entire rest of the business put together because they were now so quick to produce. If they needed a new stat and we weren't backed up on something else they could normally get it that day.<p>Cost savings? New area, new function, almost impossible to quantify. But I'm sure we saved development cost just on the initial set they'd requested, I know we were this way able to automate some reports they were having to do laboriously by hand before and if they couldn't get more sales off the back of the increased range of stats they asked for and got then I'd be disappointed in the sales &#38; marketing team.

I work for an insurance company, that handles a lot of money for what's standard in my country (Uruguay).<p>One of my coworkers found a flaw in the way the company set aside money to pay for the car repairs. He talked to the CFO, and the ensuing change ended up making the company quite a bit of money through interest (definitely much more than 75k pounds).<p>The technical change itself took all of 15 minutes (a badly written SQL select).<p>I'm still impressed at how insurance companies (and banks and other financial institutions) can still make money while being massively inefficient.<p>At least in my country's case, I believe it's a combination of a huge cost of entry and stifling legal regulations, but banks at least are having their lunch eaten on the consumer lending front by smaller competitors (still no good alternative on other services like payroll and big corp. money management which is where the big margins are I guess).

An organization I used to work for wanted to be able to analyze search keywords entered on the website's search form. But the search was a homegrown system.<p>So several business departments and the IT department entered into a series of long debates about whether to spend time adding extra functionality to the homegrown search or to just upgrade to something like Google Custom Search / Google Search Appliance -- both options that would have taken a long time, given the bureaucracy at the organization.<p>Meanwhile, in looking at the site's logfile analyses, I noticed that the analysis software they were using was really good at letting you see stats on GET queries in URLs. A quick look at the HTML for the site's search form revealed that the form was using POST.<p>I suggested to the IT dept that they simply change the search form to use GET instead of POST and call it a day. Done.

One client was charging-off sizeable bad debts on their accounts receivable, but the program did not recognize that some of the receivable balance was for the sales tax.<p>With a five-line change, $250k/yr was saved (you don't have to remit sales taxes for money you never received).

Two lines of code (and Slackware Linux) and I saved 'The Firm' around $250k (NZD)<p>Was an old Dental system written using a flavour of Unix that I've forgotten now. The system ran over X.25, but we were moving to TCP/IP. In the end I built up a bunch of slackware boxes that redirected tty1a to a telnet session and back again. Worked a treat. And the bonus was - we were then able to remote administer the original system (which wasn't possible before).<p>But while it was only two lines of code, it did take me a couple of months to figure out a solution. I had to learn a lot to get to that point - but it gave me a great appreciation for Linux.

Shortly after I started a job, I noticed that we had large spikes in our traffic stats. Updates would be sent to customers for various pieces of data on a regular basis. They have previously implemented a simple form of incremental updates, but because of the way it was implemented, it would periodically have to send all of the data again. This was happening about once day.<p>The spikes were often enough to push the 95th percentile up to an absurd amount. I spent a few days making some changes to make all updates incremental and ended up saving the company about $50k/mo in bandwidth charges.

Here is one I submitted to YC under the cool non-computer hacks: In the early days of ppc advertising when it was straight auctions. I would call up all our competitors and have them bid just 1 cent more than the previous guy so they wouldnt loose their place but dropped our cost per click to 1/4 what we were paying before (we were ranked first). I think Google changed there bidding system because of me :)<p>Another one was where we replaced a site created by March 1st for over 1 million dollars with a Yahoo! store in 48 hours for about $50.00/month

About 10 years ago I worked at a company called Nu-kote that did printer / ink supply chain stuff. They had clients like office depot office max etc... There was a system alled extenterprise that was like the daily dashboard / forecasting system for the sales team. On my 3rd day I fixed some problems that extenterprise was having. After about 3 months I was given a bonus because my forecasting algorithm was substantially more accurate than what was previously being used and it saved / made the company tons of money.

Yep. I quit. (and launched a start-up)

In one of my college jobs, I was the only software-knowledgeable person working in a test lab full of hardware people.<p>Their test-tracking and defect-tracking process required essentially reading data off of a machine and re-entering it; I figured out how to scrape the data from the terminal and use GUI automation to populate the appropriate fields at the click of button. This saved several hours per person per week over 40-50 people.

I spent a week writing an Excel macro for my old boss (hardware engineers tend to use Excel for everything!) that did literally a full day's worth of tedious work in about one second.<p>I never did get an estimate on the savings, but he uses this macro whenever there's a problem in manufacturing. So that's one senior engineer's day of work, plus one day of manufacturing doing nothing, saved each time.

I was hired to do consulting so it wasn't my direct employer but this was a pretty quick change (&#60;5mins) that saved a lot of money(&#62;$500,000 / year):<p><a href="http://analytics.blogspot.com/2010/07/using-wrong-tracking-code-can-cost-you.html" rel="nofollow">http://analytics.blogspot.com/2010/07/using-wrong-tracking-c...</a>

Thats very interesting, many many years ago I almost built a startup on your very post - but changed my mind at the last minute. Basically workstation software that reported back to a central server its idleness in increments of X minutes, giving large orgs an overall picture of their idle machine drain.

In a retail environment, I instilled more discipline in the workflow so that committed timelines were met thus improving customer service. Also started up- and cross-selling services.<p>Sales improved by 42% in one year.

I've saved a number of people quite a bit of server costs with memcached.

The bigger the company, the easier it is to do this. I try to track major things that either avoid expenditure or save money while improving services or eliminating useless spending.<p>Generally, I'm able to hit 3-4x salary. Once I was able to save the place 10-15x my salary by aggressively decommissioning some ancient, expensive to maintain legacy applications. Of course, this work is usually a team effort... but it's still a good feeling.