You don't learn from mistakes if it doesn't hurt.

Two days ago, I made a mistake by buying an item that had great margins, but which turned out to be a short-term spike in prices. I hadn't looked at the historical prices before buying. I got lucky, and still managed to sell the items before the price dropped below what I paid for it.

I repeated that same mistake yesterday, but this time, I did lose money.

I was doing some research to find deals, and quickly landed on a item promising 30,000,000 return. I got greedy, and decided to go for a quick sale, rather than do my due diligence. I flew 5 systems out, picked up the lot, and flew it back to Jita. Just in time for the Jita price to collapse. At this point, I'm looking at a 4,000,000 to 5,000,000 loss.

What did I do wrong?

1. I did not pay attention to historical price. The selling price for this item was 5,600,000 at the time I purchased the initial lot. However, this item normally trades at 4,800,000, which is also the price I paid. 
2. I did not look at the profit margin, only the net absolute profit. The margin was a paltry 11% (after broker's fees and sales tax), which doesn't give me much room to manoeuvre if I get into a pricing war.
3. I invested about 1/3 of my total cash in this lot. A fair chuck of my ISK is now tied up in goods that aren't selling.
4. I didn't look at volume. 75-150 units of this item move a day in The Forge region, with the occasional spike in the 300 range. My lot represents a fair portion of that, so it's not likely that I'll be able to sell it quickly. 

Lessons Learned

1. Always check historical prices to see if you're not dealing with an item that's seeing a temporary spike in sales. If the spike is upward, don't invest. If the spike is downward, then it may be worth spending money on it. I think a good checklist of things to verify would be a good thing.
2. Always calculate margins, to see how much room you have to adjust prices. This is the margin of security concept real world investors like Warren Buffett talk about, and it's a rule I've very familiar with. I totally ignored this rule yesterday.
3. Diversify. I was reading Full Disclosure, and saw that Vincent has a large number of items for sale. I think this could help offset losses if you're wrong about a deal or if you run in a one-time negative event. Also, you don't have all your cash tied up in one item. I don't know what the proper limit is though. Maximum 5% available cash? Maximum 10% if it's a really good deal? 
4. Always compare historical volume to current volume on the market.

I also need to figure out how to deal with under-cutters. It's very draining to have to reset prices constantly.