Gold Farming Research – Part Three

December 12th, 2009 in Buzz,WoW Gold Farming

Economic Analysis of Gold Farming

“Economics sees value wherever humans decide that some construct of theirs has utility but is scarce. Synthetic world goods have utility and are scarce; thus they have value that can be measured in terms of real dollars.” (Castronova 2006:52)

The supply—demand economics of gold farming are therefore very simple. Some people in the world have more money than time. Other people in the world have more time than money. The former demand finds the latter supply via various physical and virtual channels. And thus a market and then an industrial sub-sector are made.

Box 4: The Economic Case Against Gold Farming

He equates those who want RMT removed with those who have had their game experience reduced by RMT. That may well not be the case because (see Section D4) a number of the former may be making moral rather than experiential judgements. (For example, one can be against the death penalty without having any direct relevant experience.) Thus the in-game utility of gold farming controls may be less than he predicts.
Adverse in-game experiences may be wrongly associated with gold farming. Bots and the like may be set up by regular players. Likewise, some of the “gold farmer killing” videos on YouTube have associated commentaries arguing they are cases of mistaken identity (e.g. Odinssword 2007). Controls on gold farming, then, will not alter these wrongly-associated adverse experiences.
For those players with vigilante tendencies, “gold farmer” harassment and killing have been enjoyable parts of their gameplay. Controls on gold farming will reduce their utility.
Controls on gold farming may introduce other disutilities for regular players. In 2008, Runescape forums were flooded with players complaining about what they perceived as the deterioration in gameplay since anti-gold-farmer controls were introduced by Jagex at the end of 2007.
Actual supply—demand curves for play utility and even gold farming are unknown (to be fair, Castronova does acknowledge this when coming up with his guesstimate that real-money trading costs US$1.50 per user per month).
Transfer of surplus from gold farmers to regular players by introduction of controls is portrayed not just in economic cost-benefit terms but in “merit” (i.e. moral) terms as a transfer of surplus from rule-breakers to rule-followers. But one can counter-argue the morality by perceiving this, instead, as a transfer of surplus from poor developing country workers who might otherwise be unemployed to rich industrialised country citizens. (Could one even argue that if the latter decide to spend less time playing, the more time they could invest in more socially and economically productive activity!)

In sum the economic case for controls on gold farming may be far weaker than Castronova argues.

Enterprise Costs and Profits

From 2005, He (2005) reports 100 World of Warcraft gold would sell to players for US$10. Of that the wages paid to the gold farmer would be US$1.40. There was thus a more-than-600% income on top of wages to be divided between the gold-farming firm and the broker. Based on the other figures given here, the broker would probably have gotten the lion’s share of this.
From 2006, a Chinese entrepreneur with “over 40″ staff reported doing US$11,500 of business each month. If we take 40 as the number of staff, that means per capita productivity of US$3,450 per year. Average staff salaries were US$140 per month. That suggests a 100% overhead on salary costs to cover other costs plus any profits (Honge 2006).
From 2006, Yee (2006) reports 1,000 World of Warcraft gold would sell to players for US$66, of which US$25 would be paid to the gold farming organisation.
From 2007, Dibbell (2007) reports a 10-worker Chinese firm turning over US$80,000 per year. In that firm, 100 World of Warcraft gold nets the farmer US$1.25, the firm then sells it on for $3.00, and it is finally sold online for US$8.00.

These figures can be summarised as shown in Table 1.

Table 1: Division of Income in the Gold Farming Value Chain
Source	Playbourer	Gold-Farming Firm	Broker
He(2005)	14%	86%
Honge(2006)	50%	50%	n/a
Yee(2006)	38%		62%
Dibbell(2007)	15%	38%	47%

On this basis we can say that, at a minimum, overhead costs and any profits add a further 100% on top of the wage bill (or, put another way, that wages make up a maximum of 50% of revenue). A more typical formulation, though, would be that – from a broker-mediated sale – half would go to the broker, one-third would go to the gold-farming enterprise, and one-sixth would be paid to the workers (though, see below, this may not true from 2008 onwards).

Zhe (2006) provides a monthly income and expenditure balance sheet for a Beijing-based firm of 10 employees plus one manager, working 12 hours per day farming World of Warcraft gold in 2005/6 (see Table 5). It assumes a productivity of 200 gold per person per 12-hour shift (i.e. 60,000 gold produced per month), selling for US$17 per 100.¹

Table 2: Sample Monthly Expenditure/Income for a Chinese Gold-Farming Firm
Expenditure Item	Expenditure(US$)	Income Item	Income(US$)
Worker wages	1,205	Gold sales	10,200
Food	360
Rent	360
US proxy server rental	300
Manager wages	180
Game subscriptions	150
Internet connection fees	145
Electricity	90
145
Total	2,790	Total	10,200
Profit		Total	7.410

These figures reinforce findings from Table 4 just above. Wages are 50% of total costs, and would be 50% of revenues if the gold-farming firm were just about breaking even; as the enterprise described by Honge (2006) probably was. However, the enterprise cited in Table 5 is not merely breaking even. Instead it is making a profit.

More than that, these figures indicate the super-profitability of gold-farming, at least during its initial years. Profits are 265% of operating costs. Zhe (2006) notes that these would be shared between the gold-farming firm and its overseas contacts. The latter would have some operating costs. Similarly, brokers will also have their costs, eating into the large share of revenue that Table 4 indicates they take. Nonetheless that Table too suggested the potential for super-profits.

That certainly chimes with some of the stories that hit the headlines. One US-based owner of Asian gold farms reported earning up to US$720,000 per year (Lee 2005). One Korean entrepreneur caught by the Korean tax service had set up a gold farming operation for Lineage II, using Chinese gold farmers, in 2003. During the four years from 2003 to 2006, he made US$9.6m (Davis 2007c).
Finally, we can also use Zhe’s figures to calculate that, all other things being equal², the enterprise would no longer be profitable if the World of Warcraft gold price fell below US$4.7 per 100 gold. As discussed next, that is precisely what happened some time in 2007.

Exchange Rates and Enterprise Viability

Exchange rates between real-world and virtual currencies impact neither regular players nor game companies in subscription-based games. They do, though, impact gold farmers. Calculations from available data on leading games show that in-game currencies, on average, devalued against the US dollar³ by roughly 75% between June 2005 and June 2008 (see Appendix 2).⁴

Box 5: In-Game Inflation, Deflation and Gold-Farming

One of players’, and game companies’, main criticisms of gold farming is that it spoils the virtual economics of game worlds (e.g. Yee 2006). Principally, the complaint is that gold farming fuels in-game inflation (e.g. Kaminski 2006, Orii 2007, Ward 2008). The economics of this are apparently straightforward: increase in the supply of any item – assuming constant demand – causes its value to fall. As gold farmers pump additional currency into the virtual economy, it is argued, this is the real-world equivalent of increasing the money supply. The value of the currency falls. It therefore requires more of the currency to purchase any item – in other words, prices rise and there is in-game inflation.

Leaving aside arguments about the real-world relationship between money supply and inflation, we can throw in a few spanners that complicate the virtual world picture (developed from Paul 2005, Yee 2005b):
1: Money supply: do gold farmers actually add to the supply of in-game currency? If gold farmers merely create currency that other players would otherwise have created, they do not provide any additional impact on the money supply. There isn’t space here to go into virtual economics in-depth but one sign that gold farmers are merely substituting one source of supply for another comes in the complaints (about “spawn-camping” and “kill-stealing”) by regular players that gold farmers are taking their currency sources. (Additionally, even if the money supply does rise, there is no simple link to prices since players may hoard⁵.)

2: Size of effect: there are few signs that gold farmers or gold buyers have become the dominant force in game economies. Thus the effect they – as opposed to regular players – have on the virtual economy will be fractional; possibly even marginal. How regular players behave economically matters more.

3: Demand: may not be constant. On the simple fact of number of players, older games are likely to see falling demand; newer games rising demand (for data, see White (2008) and Woodcock (2008)).

4: Game company effects: game companies, not gold farmers, are the game’s “central bankers” controlling the ultimate source of money/item supply and also impacting demand. Game patches and redesigns may introduce new sources of in-game currency, or new sinks; they may also increase or decrease the demand for certain items and for currency.

5: Deflationary effects: of the three main farmer activities (currency, items, power-levelling/account-selling), the last is likely to be relatively economically neutral, not least because it follows the behaviour of a regular player. In most games, gold farmers have only marginal ways to directly “farm” currency (e.g. collecting the relatively small amounts that killed monsters drop) but much more often, they will be gaining items and either selling them in-game for currency, or selling them out-game for real money. In either case, they will be increasing the supply of in-game items. Assuming constant demand, that will cause the price of those items to fall, having a deflationary effect on the virtual economy (though the caveats above apply to this also). That is certainly something about which regular players complain.

6: Data: in-game inflation has undoubtedly been seen but there appear to be few long-term records. One shorter-term set for a game in which gold farming is present, EVE Online (for October 2005 to June 2007) shows deflation, not inflation (Lehtiniemi 2008). Similarly, comparing archived and current prices for a basket of different items on the Runescape Price Guide (http://www.zybez.net/priceguide.php) indicates a mix of inflated and deflated prices but an overall deflation of 25% between September 2006 and July 2008. Yee (ibid.) also claims in-game deflation in World of Warcraft.

In sum, the linkage between gold farmer activities and in-game inflation (or other price changes) currently receives a “not proven” verdict.

Increased productivity: finding ways to make more in-game currency per hour. This seems very plausible. Typical in-game earning rates for World of Warcraft cited from earlier work are 200-300 gold per 12-hour shift (e.g. He 2005, Zhe 2006). In 2007, Blizzard provided ways for players to earn more in-game. Current gold-earning guides indicate 100 gold per hour can easily be made, with several hundred per hour being feasible for highest-level players (e.g. Voodex 2008). It is possible, then, that productivity increases might fully compensate the devaluation (and also, given supply—demand economics, that they might have helped cause that devaluation). As discussed in Section A2, we know gold-farming firms do invest staff time in researching ways of raising currency-per-hour productivity (Dibbell 2007).
Automation: making use of macros, bots, etc. to increase productivity and/or cut costs. This is possible but there is little data to support or reject it. Gold farming web sites typically claim never to use bots. But customers whose characters are being power-levelled do report use of bots on their accounts (e.g. Allen 2008, PowerLevelingReviews 2008). On the other hand, Mithra (2006) claims low-wage labour can even undercut automation since the latter does have its costs.
Revenue adjustment: changing the profile or distribution of income. The super-profitability of gold farming may have been eroded by competition/devaluation, leading to more normal or even tiny profit levels. The conceivability of this is strengthened by claims of significant profit erosion after 2006 from industry insiders (Terdiman 2007, Salyer 2007).
Disintermediation: cutting out brokers. Type “buy WoW gold” into Google in mid-2008 and you got four million hits. Sampling these suggests many appear to be gold farmers and their firms seeking to sell direct to customers. Archive analysis shows almost all started from 2005 onwards. And we know the main intermediary – IGE – lost market share significantly from 2006 onwards (Salyer 2007). From the figures given above in Table 4, removing brokers from the supply chain would significantly increase potential enterprise earnings and/or help compensate for currency devaluation. This therefore also looks like a possible explanation.
Diversification: into other activities. An obvious example could be greater emphasis on power-levelling work, the price of which will not be affected by currency devaluation. However, data from Barboza (2005), Carless (2007) and from gold farmer web sites suggests that the price to level a character from 1-60 in World of Warcraft cost very roughly US$350 in 2004, US$250 in 2006 and US$150 in 2008 (though with great variation between sites). These price falls are somewhat lower but not greatly dissimilar to, those for in-game currency, and probably equally caused by new entrants and competition. This type of diversification may thus not a particularly viable strategy to address devaluation. An alternative may be focusing on newer online games, where competition may be less and margins higher, but evidence is absent. Likewise for another potential diversification that appears on a number of gold-farmer web sites: sale of game account activation codes (“keys”) and prepaid gameplay time cards. One could also view these as economies of scope – enabling the same marketing, sales and fulfilment fixed costs to be spread over a greater number of services.

In summary, then, and on the basis of far-too-limited an evidence base, productivity increases, disintermediation and profit erosion seem the most plausible explanations for gold farmer survival in the face of devaluation of in-game currency, alongside a number of other possible reasons.

Gold-Farming Scale Economies

Economies of scale exist “where a firm can lower the cost of each unit of output by producing more units” (Sayer 1985:10); meaning that firms producing larger amounts have a competitive advantage because they can produce each item more cheaply than a smaller producer. They can do this where there are fixed costs: input costs which do not rise proportionately for each extra unit produced. The alternative is variable costs: which rise proportionately for each extra unit produced.

Basic in-game gold-farming appears to have few scale economies. Variable costs dominate as each additional production unit (i.e. individual or pair of players) requires one PC, one Internet connection, and one account. Players are typically paid incrementally based on output, so wage costs are fully variable. Productivity per worker is also constant – one person kills a monster or chops wood or crafts an amulet just as quickly whether there are not or twenty other co-workers doing the same.

gold farming research part one

Indivisible-cost items: some investments, although their input cost does vary with size of output, are discrete items (“lumpy investments”). For example, if it requires one manager to manage 20 gold farmers, or one technician to manage 20 PCs, this creates scale economies on the assumption that it is hard to purchase twentieths of their services – they only come in discontinuous amounts: zero, one, etc. There can also be an equivalent indivisibility in game-play. Some high-level monsters can only be killed by groups working together; hence the items or currency they drop have a scale economy.
Fixed-cost items: some gold-farming firms will have fixed-cost investments in a web portal, in setting up payment and security systems, and in marketing their services. These create scale economies.
Divisions/specialisation of labour: as noted above (in Section A2), gold farmers play various different roles in-game, and staff in gold-farming firms undertake different out-game activities. On the assumption that there are efficiencies gained from specialising in particular roles, then there will be scale economies for those firms that have enough workers to allow this specialisation.

The last two items on the list help explain why individual gold farmers may likely work via brokers/exchanges rather than seeking to sell direct since they thereby avoid fixed costs and the need to adopt multiple roles. If all three items in the list were economically overwhelming, then one would expect medium- and large-sized enterprises to emerge. That they probably have not suggests fixed/indivisible costs do not dominate⁷.

There may also be scale diseconomies – from perceptions that the sector is too volatile to justify large-scale investment; to dangers of “becoming noticed” e.g. for taxation purposes by local government or for legal action by game companies and/or the inability to cut regulation-related costs as informal sector small enterprises can; to growing costs of coordinating a large operation. It appears possible that the presumed-largest player during 2004-2006, IGE, suffered all these diseconomies during 2007 (PJ 2007b).
B4.

Information Failures and Real-Money Trading

In some ways it is a minor miracle that gold farming can exist as a sub-sector given that real-money trading is such a textbook case of information failure. Trading generically relies heavily on information during each of its three steps (Norton 1992, Casson 1997):
1: Information acquired prior to trading (on general items/services available, on the existence of traders, on their reputation and trustworthiness, on typical prices).
2: Information communicated during trading (on specific items or services offered and money sought, on quality of items/services offered, as part of negotiation).
3: Information acquired after trading (on whether or not the terms of the agreed trade contract have been fulfilled).
Availability, quality, cost and other characteristics of information, and the ability to communicate that information, are thus critical foundations for all trade and all enterprise (Porter and Millar 1985, Stiglitz 1988).

Given that everyone playing MMOGs and all gold farmers have web access, and given the huge quantities of data available on gold farming⁸, then information failure might, at first sight, seem odd. The key problem is at least three-fold – the virtuality of trade (buyers and sellers never meet physically), the anonymity of online activity, and data quality (the snowstorm of data available that could be good, bad or indifferent). Data available online may be good for providing buyers with certain information – the virtual existence of sellers, typical prices, specific items and services offered. But the following information failures still occur:
1: Information absence: both buyers and sellers may be completely unable to find out who, in reality, they are trading with.
2: Information uncertainty: buyers and sellers will be uncertain about each other’s trustworthiness; buyers particularly will be uncertain what – if anything – will really be delivered if they pay; buyers report being uncertain about whether or not partially-completed deals will ever be fully-completed; both sides will be uncertain about whether or not their trade is under surveillance from game companies.
3: Information asymmetry: absences and uncertainties affect both sides of real-money trading but there is a typical asymmetry since key items of information about the trustworthiness and quality of items/service are known to the sellers but not the buyers.
4: Communication problems: sellers typically offer online chat and email contacts but buyers report problems with communication relating to issues like language, time difference, non-response and being fobbed off with excuses (e.g. mmobux 2008).

Informational characteristics such as these failures in turn shape both the process and structure of commerce (Williamson 1975, Stiglitz 1988). The informational characteristics just described mean the build-up to trading may be relatively quick and easy. However, trading overall has the characteristic that it is risky; far more so than normal real-world trading.

That risk can be instantiated as both opportunism and adverse selection. Opportunism would mean actions such as overcharging for goods or agreeing to a contract knowing it cannot properly be fulfilled. One can seek evidence for this from those who post online about the experience of buying gold-farmed items/services. They are generally negative (e.g. Jamie 2007, Allen 2008, PowerLevelingReviews 2008). Of course one must allow for the profile of those who post being different from the average buyer profile, and the possibility that posts are made by those with vested interests for and against gold farming or particular suppliers. However, the level of detail provided in some posts suggests they represent real experiences, and that a proportion of purchasers are disappointed. Examples include:

Late delivery: rather than the instant service and large stock promised, purchasers find currency being delivered piecemeal over a long period of time; other actions promised quickly do not occur for days or even weeks.
Partial delivery: full amounts of currency are not delivered; characters are returned having been only partially-levelled.
Currency loss: currency is impounded by the game company.
Account suspensions and banning: particularly for power-levelling.
Disputes: as purchasers try to get their money back.

Underfulfillment and opportunism thus do seem to be present.

Adverse selection would mean actions such as unwittingly selecting either a trade partner or trade items of poor quality. The quality of virtual trade items can readily be determined on their delivery but the risks of poor trade partner selection do appear to be present. They are present for sellers (e.g. defraud by players: see Box 12). And they are present for buyers. For example, out of more than 400 real-money traders reviewed by mmobux⁹, only five got a rating of more than 7 out of 10, sufficient for them to be deemed “extremely reliable” (Carebear 2008). The vast majority of traders got very low ratings indicating a poor quality of trade.

Suppression of trade: the level of trading is likely to be below that which would occur if the various informational challenges were mitigated or removed. One possible indicator – already mentioned above – is the gap between the US$10 average annual spend per player ordinarily and the US$46 spent when Sony set up Station Exchange for Everquest II; a system that addressed most of the information failures indicated here. That suggests a possible 78% suppression of trade due to information failures (though issues of legitimacy and effort would also have an effect).
Localisation of trading: as traders seek to deal only with those they physically know. As indicated below in Section C1, RMT did begin on this basis. It still seems to be the starting point for individual developing countries as MMOGs take off. However, the online nature of the games, including their globalisation, has encouraged trade to move beyond the local, impelled by buyers and sellers not knowing enough other players to match their respective demand and supply.
Presence of intermediaries: intermediaries address information absences and uncertainties by holding information about both buyers and sellers; for example, reputational/trustworthiness and quality information. They can reduce the information-gathering costs of all stages of trading. They can make trade less risky, or at least make it perceived to be less risky, because of their informational resources and reputation. In practice, though, the emergence of intermediaries (e.g. brokers or exchanges) appears relatively limited¹⁰. Pressures for disintermediation thanks to the virtual nature of trade, and pressures from game companies are a partial explanation. The shifting and anonymous nature of buyers is another. And the brokers – though less so the exchanges – have their own reputational problems (e.g. PJ 2007).
Reputational portals: given the importance and scarcity of information on reputation and trust, it should have a high value, and this should encourage information brokers to emerge who would gather and disseminate such information. In practice, there appear to be relatively few such brokers, most likely because they also struggle to establish their own trustworthiness and that of the information they provide. The one exception appears to be: http://www.mmobux.com^11. Exchanges of which, again, there appear to be relatively few (e.g. http://www.markeedragon.com, http://www.playerauctions.com/, http://www.randyrun.com; http://www.worldmmo.com; http://www.gamexa.com/), typically provide this information as an integral part of their service (though the latter two appear to have only a few suppliers). (Comparative web traffic information on these exchanges can be found in Appendix 3.)
Reputational tactics by sellers: given information uncertainties and the importance of trust, sellers would be expected to try to provide a lot of information about their reputation and trustworthiness. Overt tactics include the presence of customer testimonials (e.g. http://www.wowpowerlevelingpro.com/index.aspx); graphics of reputed global firms such as Mastercard, Visa, PayPal (e.g. http://www.gmlvl.com/); links to reputation rating sites such as BizRate (e.g. http://www.igegolds.com/); guarantees of fast, safe service and refunds if unsatisfied (e.g. http://www.mmoempire.com/); and demonstrations of altruism/corporate responsibility through donation programmes to Western charities (e.g. http://www.ige.cc). Other tactics include advertising methods for easier communication such as live chat; detailed explanations of the process (e.g. http://www.mmorpgrealm.com/); and imitation of the names of well-known players (e.g. http://www.ige.cc). The main problem is that, by and large, all these tactics are perceptual rather than real. They may have a marginal impact on some customers but they have no more actual value than a would-be real trader telling you “I am not a fraudster”.¹² One should also mention the potential for “anti-tactics”: providing false negative information about competitors. This means that even where companies are rated – for example, IGE has more than 25,000 ratings on BizRate – it is difficult to trust either the positive or the negative ratings.
Repeat business: those purchasers who buy repeatedly are likely to stick with one supplier if they are satisfied with its service. Given the lack of data on purchasers, it is not clear if this happens in practice.

From all this, it can be seen that the informational characteristics of gold farming and real-money trading do, indeed, shape the activity in a powerful manner. For example, they probably significantly suppress the level of trade. However, because of virtuality, the outcome is not exactly that which one would predict from real-world experiences. In particular, trade is more globalised and seems less intermediated than an offline equivalent with similar information characteristics (of which the drugs trade might be one example: see Box 16).

Box 6: Open and Closed Economies

James Clarke, Chief Operating Officer for the gold/services broker IGE, has made a real-world analogy, claiming that RMT is good for online games because closed economies are “not usually as vibrant” as those that are open to imports and exports (Carless 2006). Of course, in some ways, the analogy is incorrect – all game economies are closed because nothing can be introduced from outside the virtual world, and World of Warcraft is not about to start importing items from Final Fantasy XI (though don’t bet against world-to-world trading, transport and battles as a next-ten-years development). Nonetheless, it might be intriguing to look at some of the macroeconomic analogies: if RMT is the equivalent of trade liberalisation, then one might predict winners and losers: those with good skills, technology and capital to benefit most; and the poorest to lose out.

1: The source of the figures is not stated though the writer does appear to have some “insider” familiarity with gold-farming operations.
2: As noted above, there is no evidence at present that wages have changed in recent years. The same is true of other costs (though one might imagine Internet connection fees to have fallen). Zhe’s calculations do not include fixed costs for hardware and software. Nor do they include IT operating costs (for instance, Mithra (2006) comments, “When you run two dozen computers night and day, even in a temperature controlled room, you end up replacing $60 this and $100 that at least once a week.”). That would increase the costs and reduce the profits shown. So, too, would the yuan’s devaluation from 8.28 to the US dollar throughout 1995-2005 to around 6.9 to the US dollar in mid-2008 following its 2005 unpegging.
3: Yet another missing piece of the data puzzle is exchange rates against other currencies. Lim (2007), for example, reports prices of in-game currencies in RMB yuan to be around one-third the price charged in dollars.
4: There are claims that anti-gold-farming campaigns by game companies do temporarily revalue currencies (Dibbell 2007). The only offered explanation for the devaluation is competition, with new entrants undercutting existing firms in order to try to win business (e.g. Carless 2007, Debatty 2008). The devaluation seems unlikely to be the result of in-game effects (see Box 5) since these point in different directions, and show no consistent relation to the external exchange rate. Note that virtual-world exchange rates do not work in quite the same way as ordinary real-world rates. For most MMOGs, it is much harder to sell currency than to buy it; and the spread is huge – US$0.30-0.80 per 100 World of Warcraft gold when you sell; c.US$2 when you buy in mid-2008. In that sense, the closer real-world equivalent is black market trade in a non-convertible currency.
5: In early 2008, players discovered that Blizzard had set a maximum limit of roughly 214,000 gold per player in World of Warcraft (Himes 2008). This is an astronomical amount for regular players and only discovered by players hoarding/not spending.
6: The, perhaps naïve, assumption is that the key and card codes provided are legal. The consistency of pricing across gold-farmer sites, and the consistency with non-farmer sites such as Amazon, suggests this is the case.
7: That would be the conclusion drawn from the figures in Table 5. However, we noted those figures failed to include hardware/operating software costs; and also excluded all costs associated with selling to end-consumers.
8: As cheap examples, Google produces 1.1m results for a search on “gold farming”, 5.3m results for a search on ‘warcraft gold’.
9: This site provides what appears to be the most comprehensive review of gold-farming firms: http://www.mmobux.com/shops.
10: This is not easy to judge – separating intermediaries from end-producer gold farms on the basis of just their web sites is hard; judging volume of trade is even harder. But the web traffic volume of the only certain broker, IGE, is no greater than that of a number of gold-farming firms; and traffic volume for the few identifiable exchanges is typically quite a bit lower (see Appendix 3).
11: Problems with other reputation-and-review sites include the following. They make their money from links to real-money traders and all reviews are positive (e.g. http://buywowgold.co.uk/; http://www.warcraftgoldreviews.com/gold_seller_reviews.php). They relate largely to one individual’s experiences (e.g. http://www.powerlevelingreviews.com/). They are very limited in coverage (e.g. http://powerlevelingreviews.wordpress.com/.) Some combination of the above (e.g. http://wowgoldbuyer.com/).
12: Just to give one example of the potential gap between perception and reality, thsale.com’s web site assures customers of the safety of its procedures and states, “We have never been informed of any account bans of thsale customers due to their dealings with thsale.” They have clearly not looked at their customer reviews on BizRate.

Related posts:

Share/Bookmark/Discuss:

Comments Off