The middleman report - a semi-product business case

[counting]

Once upon a time, in the land of Capitalism Lab world, People say it's a land of opportunity, but you can't survive for conducting a kind of business called - B2B. The pretext seems normal, all your business relationships have to be with AIs both upstream and downstream. But I added some additional specific rules, you can't touch farming and raw materials productions, nor touch the retail businesses. And of course no media firms boost, nor real estate rent income. Manipulate stock market is the big no-no. How do you survive in this situation with AIs surround you from all sides.

The game settings are : Difficulty rating 203%. And the rest as follow

setting1.jpg (70.87 KiB) Viewed 3993 times

setting2.jpg (81.51 KiB) Viewed 3993 times

setting3.jpg (50.37 KiB) Viewed 3993 times

And let the middleman challenge begins.

First, I use just 1 R&D to research 6 semi-product techs. The company's future lives or dies all depend on it. Since from experience I know electronic components have decent profit margin, and it's a crucial ingredient so a lot of potential buyers. Hence I choose it to be my first main product, and put more units in R&D to research it faster.

RD start.jpg (168.63 KiB) Viewed 3993 times

At year 5, the tech is ready, so the first "high tech" electronic components rolling off the factory and start our middleman business. The basic setup is as followed : 1 medium warehouse for source materials insurance, 1 large warehouse as wholesale portable for electronic components, 3 large factories for producing electronic components and intermediate semi-products. The AI clients flooded to my high quality wholesaler warehouse, even with seaport product as my competitors. More than 80% of the market share of electronic components is mine.

client list and market share.png (159.98 KiB) Viewed 3993 times

After 2 years of operation and 7 years into the game, the revenue reach 85m, and net profit 30m

profit 2 years.jpg (179 KiB) Viewed 3993 times

After 10 years operation, I step into CPU, and CCD semi-product as well, and I've added another R&D center to research these products. I'v controlled 50% of both global CCD and CPU market.

CCD.jpg (191.39 KiB) Viewed 3993 times

CPU.jpg (207.11 KiB) Viewed 3993 times

With the additional large factory to produce them, and I began to self sustaining steel and silicon, thus also take a small chunk of steel market share. I've double my revenue to 167m, and increase net profit to 50m, at 16th year.

profit12.jpg (178.29 KiB) Viewed 3982 times

At the peak of the businesses my revenue once reach 240m, and nearly 90m profit. However, after more than 20 years, AI finally decided to step up it's R&D and compete with my electronic component market. so my profit shrink a little bit and stay around 70m. The peak save file is as attached.

Middleman.rar

(6.65 MiB) Downloaded 1004 times

[David]

I think you are one of the few players who have had successes with B2B business in CapLab, as it is far more challenging and difficult to run a viable B2B business than a traditional B2C one.

[counting]

I'll link my original reply here
http://www.capitalismlab.com/forum/view ... t=30#p9032
And my thought about how to implement smarter AIs as semi-product producers as well as other risk-involved businesses and allow cooperation.
http://www.capitalismlab.com/forum/view ... =559#p9033
I believe mwyeoh also said in an old post about he's able to run semi-product businesses as well, although not to sure how he did, and what he concerned as semi-products (including farm semi-products and raw materials perhaps? They are way easier than mineral based semi-products)
http://www.capitalismlab.com/forum/view ... 320&p=1207

[jondonnis]

It's tough. I got stung with a recession just after r&d of the ciruits. Went on for a bit but then an ai sold them. Crap ones but cheap so everyone bought his. I put mine to £3.20 roughly giving me 115 rating forcing him to go as low as 99cents. But he just accepted the losses caused he was making money elsewhere.

I didn't bother with warehouse and sold direct from factory. Didn't understand the need for the warehouse? Tried to sell the steel which sold bit of but couldn't sell my silicon even though was the best.

My company wasn't showing on the dominance list at all for some reason. Are circuits etc not part of that list? How do you know if you're 1st then?

Eventually went bankrupt. 100mill loan at beginning probably doesn't help due to the interest. Maybe I'll issue shares next time to fund it all.

Despite being in red most months my share price was higher than any other company which I didn't understand.

[counting]

It's tough. I got stung with a recession just after r&d of the ciruits. Went on for a bit but then an ai sold them. Crap ones but cheap so everyone bought his. I put mine to £3.20 roughly giving me 115 rating forcing him to go as low as 99cents. But he just accepted the losses caused he was making money elsewhere.

I didn't bother with warehouse and sold direct from factory. Didn't understand the need for the warehouse? Tried to sell the steel which sold bit of but couldn't sell my silicon even though was the best.

My company wasn't showing on the dominance list at all for some reason. Are circuits etc not part of that list? How do you know if you're 1st then?

Eventually went bankrupt. 100mill loan at beginning probably doesn't help due to the interest. Maybe I'll issue shares next time to fund it all.

Despite being in red most months my share price was higher than any other company which I didn't understand.

The wholesale warehouse is to minimize fluctuation. Since semi-product purchasing is kinda "seasonal" (AI buys them in a batch and wait for quite a while to purchase the next batch), so you need to maintain a much larger "client lists" that at any given moment some of them are buying. Like here, my 21 clients of electronic circuits, on average about 4 or 5 were active, but at low it could be 2 or 3, at high it could be half of them.

If you don't use a warehouse that always have supply over demand and use a factory instead (ofc you could still do it with factories that have inventory units, but it would sacrifice manufacturing capacity, and wouldn't work when multiple factories are needed to satisfied demand), when a client go "online" to purchase and currently it's hot season where a lot of clients are demanding, this new online-AI will see your factory doesn't have enough supply at hand, thus try to find another source if possible. And since there's a "grace period" and "price range" that AI would stick to a supplier, once an AI client is lost, it's very difficult to get it back. However on the positive side, if you can keep it in your client list with ample supplies, they won't switch to another suppliers that easy, hence you don't always need to match your competition's rating. As long as the rating is within acceptable range (this range is the tricky part that required some experience to find it), you can set the rating a bit lower than competitors and clients would stay. This would also prevent a price war with cheap low quality competing semi-products.

This is called a challenge for a good reason, require certain understandings of AI purchasing behaviors.

[counting]

It's tough. I got stung with a recession just after r&d of the ciruits. Went on for a bit but then an ai sold them. Crap ones but cheap so everyone bought his. I put mine to £3.20 roughly giving me 115 rating forcing him to go as low as 99cents. But he just accepted the losses caused he was making money elsewhere.

I didn't bother with warehouse and sold direct from factory. Didn't understand the need for the warehouse? Tried to sell the steel which sold bit of but couldn't sell my silicon even though was the best.

My company wasn't showing on the dominance list at all for some reason. Are circuits etc not part of that list? How do you know if you're 1st then?

Eventually went bankrupt. 100mill loan at beginning probably doesn't help due to the interest. Maybe I'll issue shares next time to fund it all.

Despite being in red most months my share price was higher than any other company which I didn't understand.

BTW since semi-products normally have ridiculous low raw material to wholesale price ratio. Your inventory could be quite "inflated" if you can't sell them "smoothly". Hence you "assets" are growing at quite a high rate from the beginning and reflected on your stock performance, the negative profit would take a while to manifest when your cash position actually start to fall. This is similar to a Ponzi mechanism, where you buy source cheap and inflated the inventory without selling them, as long as the overhead is small enough over time, the game of "cooking book" and inflated stock performance can last a while.

For a particular semi-product dominance, just check the pie chart (as shown in my post) or in products page filtered with semi-products. If you really are dominance it's hard to miss (all cities share the same pie chart portion, each city only takes a partial of the pie chart. Totally add up to 100%). And semi-products don't show up in score or dominance page, which by the way should be patched. Maybe you could report this bug.

[jondonnis]

That explains why my shares were so high then thanks.

Taking your tips and trying again. This time with a warehouse

[jondonnis]

I know you set full training on the R&D centre. Did you set any training on the warehouse and factory at the start?

[jondonnis]

Going OK so far. It's difficult avoiding the temptation to put a mine down to get better quality resources

[counting]

I know you set full training on the R&D centre. Did you set any training on the warehouse and factory at the start?

I usually set training budget for factories depend on future demand expected to reach. The tricky part is how to determine the expect demand quantity? Generally I would start of with factories with minimum training and test run for about a year. Then find the highest demand month aggregate quantity and use it as a guideline. And since new tech level will bring down manufacturing level, I would later set the budget to a level that would maintain at a certain manufacturing level that could keep up with that expected demand. (P.S. Notice that I left out a lot of the calculation, since it would take pages to explain, if you don't try to do the math, you can eyeball the demand/supply bar changes over the months, and tune it manually. Just remember try to keep supply higher than demand, the moment you find there is a upward demand increase, increase the training budget, or even open a new factory).

What would normally happened for me is that I usually over produce in the beginning to fill up warehouse, hence later when I am about to up the training and warehouse is more or less stable with stockpile, some of the factories will be idle. At that point I would convert them to other semi-products with just few losses on training, since convert to other product will erase all training on manufacturing and drop a level for input and output (as in this particular report, I converted one of the factories to a CPU/CCD hybrid factory)

For warehouse training, it's always depend on total demand. Normally I would up the training to half in the first year and see if supply can keep up with demand, and then simply tune it afterward depends on the utilization. At this particular report, I eventually tune it up so it's output units can reach level 9 to fulfill all the demand. The good thing is that warehouse trained pretty fast when they have high utilization. For the input buffer warehouse, usually minimum training is enough, since upstream throughput is usually quite low, no point training up.