LindeX Selling Strategies

We’ve got to start selling L$ on LindeX. There have been a number of holdups until now, mostly related to the spiraling LindeX rate and panicky market. But the exchange has been stable now for almost a month. That stability could be due to a lot of reasons – Greenspanian pronouncements, reduced sources, transparency of economic info, etc. – but I think the biggest reason is just what Peter always said: We finally have buy limit orders to normalize activity, and now that they’ve had a chance to bake in, the market is stable.

In any case, the market is stable, and we’ve said that once the market is stable, we’d start selling L$. Sure, it will panic the market again, but if we do it right, we have enabled an incredibly high-margin revenue stream. So how do we sell? How much do we sell? At what price?

Quite a few folks argue that the amount sold must correlate exactly to some L$ charge, like upload fees or classified ads. I think the idea behind that thought is that the market will accept L$ sales that translate directly to services that the user community believes is beneficial, or at least believes has a direct relationship to system resources. I disagree with this for at least three reasons:

1) The market doesn’t think clearly enough, or monolithically enough, to accept that any particular L$ sink is a good candidate for L$ sale.

2) There’s actually no rational relationship between the amount of L$ we can or should sell and any one particular sink.

3) Someday – as we reduce the free sources of L$ – we need to be selling 100% of our sinks, and eventually even more than 100% our sinks. (It’s tautological: When there are no free sources, new L$ can only go into the L$ economy through sale of L$.) So we shouldn’t let the market get into a mode of judging whether any sinks should be sold, since someday we will need to sell all of them, regardless of whether the market thinks they’re “worthy” of sale.

So then, here is the selling strategy I’d propose instead:

Make an announcement on the forums to this effect: “At noon tomorrow, Linden Lab will place a sell order on LindeX in a US$ amount equal to [X]% of the L$ sinks over the last 30 days (including today). At the average sale price over that same period, this would be about US$[Y]. The offered sale price for this block will be the average sale price over the last 5 days, ending with today, or approximately L$[Z]. At the same time, we will place a buy order for exactly the same US$ amount, at the highest selling price for which there has been more than one sale over the last 365 days, currently L$[361? Peter to confirm]. After this sell and purchase order, Linden Lab will not buy or sell on LindeX for a minimum of 14 days, at which time we will evaluate whether the Linden economy requires additional L$ infusion.”

In other words, “We’re selling Linden Dollars. It’s a small amount of the last month’s sinks. If you want to panic, go ahead and panic. But if you panic through the highest price of the year, you’ll be selling right through our buy block that is the same dollar amount we just sold, so it’s like nothing happened.” Of course, the market could panic right beyond the block, but there’s no logical reason to under almost any theory. And since there’s no reason to panic through the high block, the market probably won’t panic at all. Well, let’s not say “probably,” let’s say “shouldn’t . . . but might anyway.”

I’d target the X% so that the $Y sale amount equals about $5000. And to the extent this proves a stable selling strategy, we’d slowly move that X% up. And eventually move to automated sales every week, then every day. Eventually, once the market is accustomed to the activity, we won’t need to put up the buy order block at all.

There are many many variations that could be as good or better selling strategies. Please feel free to comment away . . .

6 thoughts on “LindeX Selling Strategies

  1. We shouldn’t contemplate selling additional L$ until we’re getting market value for the L$ we’re *already* selling.

    Let us assume that 50% of the premium fee is for the land, and 50% is for the stipends. In May we had $179K in revenue for premium accounts, so that’s $89,500 for the stipends. In May we also gave out approximately L$51M in premium stipends, which means, in effect, that we’re selling at L$570/$1! (Remember that many premium users are paying less than $9.95/mo through quarterly and annual subs.) If we had sold the same amount at the current rate of L$330/$1, we would have brought in $154K. We left $65,000 on the table in May alone.

    For a while we’ve been talking about decoupling land and stipends, and (probably) turning the “stipend” into a scheduled market buy, thereby creating demand that we can legitimately sell into at prevailing rates. The plan is not without downsides; some people will be unhappy that their monthly $9.95 doesn’t go as far as it once did. Nonetheless, we must eliminate the enormous disparity between stipends and the market, and we must regain control of the number of L$ we create each month.

    Once we’re selling at fair market value, then we can have the debate over how much to sell, and when, and how. There has already been extensive discussion on this topic:
    https://jira.lindenlab.com/browse/SL-6201

  2. I strongly believe that we should NOT begin to sell L$ until we have eliminated Premium Stipends (which requires that we have scheduled L$ buys & inventory limits in place).

    We can make a credible argument that we need to sell L$ to support the liquidity needs of an expanding economy and population ONLY if we are not already expanding the mL$ supply through massive stipends. Selling L$ without doing away with stipends communicates the message that we are reaching for more revenues at the expense of our residents.

    We should also be careful not to underestimate the impact of generating panic in the currency market. If we cause a further large deterioration of the value of the L$, we should expect that our Premium Attrition rates will increase significantly. If a big currency panic occurs, the loss of revenue from high Premium Attrition would more than offset anything we made by selling L$ on the exchange.

    The other key step that we should take before we start selling L$ is to publish a credible L$ management policy. Once we have eliminated Premium stipends, we should publish a policy document that describes how we will go about determining how many L$ will be placed on sale each month. To my mind, the key factors that would drive this L$ supply policy would be:

    – L$ sales should be based on the observed growth of in demand for currency. The growth in demand results from both an expanding active user base and changes in overall spending per user.

    – Long-term pricing stability should also be a key goal of our L$ management policy. While we should not strive to keep the L$ in a specific trading range, long term trends in the value of the L$ must be factored into our decision making process.

    I think that we absolutely should move towards selling L$ in the not-too-distant future. However, we need to tread very carefully in this area. We have seen the value of the L$ is currently driven far more by psychology than fundamental shifts in the economy. This suggests that we will only be successful in turning L$ sales into a significant revenue source if we can convince most of the community that our actions are beneficial (rather than harmful) to the overall health of the SL economy.

  3. [rant]

    Not selling lindens on the exchange ASAP is fundamentally a bad business decision no matter which way you slice it. Sure, once we announce that we’re doing this a vocal portion of residents will cry ‘Jihad!’. Who cares? That’s their default reaction to any change. Trust me…they’ ll be jihading against something else next week.

    Our core business model isn’t based on Estate Tools or Voice Chat or other nice-to-have features. It’s based on revenues, and the ability to realize revenues collected from micro-payments is a critical component of our goal towards cash flow break-even. It’s a HUGE mistake to equally weight selling L$ and resident tools as if they’re interchangeable features.

    We need to butch-up and move forward with selling L$ on the LindeX. Our endless economic theorizing and community backlash discussions are unproductive, ineffective and expensive. Failing to properly prioritize this issue and execute on it has the same effect as consciously deciding to miss our revenue targets for 2006 and possibly 2007.

    [/rant]

  4. Ian makes a good argument, but I’d like to clarify the calculation and also give an alternative perspective.

    First, on the calculation of fees attributable to stipend: I’d peg it at US$5 per premium account that month rather than 50% of the revenue basis. Remember, “revenue” reported in https://osiris.lindenlab.com/reports/index.php?section_id=161 is neither revenue on an accrual basis nor cash receipts. Instead, it is “fees assessed net of credits and bad debt.” This differs from accrual revenue in that it counts the charges assessed (whether collected from new cash inflow or application of account credit) that month, and not the portion of the charge recognizable as revenue that month. So, for example, “revenue basis” counts the $72 assessed on an annual premium account that month, even though we would only recognize $6 of revenue from that charge. Because most of our charges are assessed in advance, our revenue basis is always higher than our actual recognized revenue.

    I would argue that we should simply count $5 per premium account to determine how much of the $10 monthly premium fee is attributable to L$ stipend. This is a bit of cheat, in that I am assuming that the value of the included 512 sqm tier is the other $5. That free tier may have some other dollar value amount, and it won’t really be definable until basics buy land. But it’s one of the benefits of basics buying land that we have a definite value attributable to the first 512 sqm tier. See this comment https://jira.lindenlab.com/browse/SL-17190#action_33337, which resulted in this task https://jira.lindenlab.com/browse/SL-17793.

    So, returning to Ian’s argument: we had 18,491 premium accounts in May according to this report https://osiris.lindenlab.com/reports/index.php?section_id=81&reports=17, and 45,469,050 in recurring stipends to premiums according to this report https://osiris.lindenlab.com/reports/index.php?section_id=60&reports=562 (Ian had a higher number because he included non-stipend activity). So by my logic, that’s 45,469,050/$92,455 = L$492/US$1.

    A little lower than Ian’s number, but still a higher rate than current market, so his argument is still good. But my perspective on this is that Premium memberships are a good deal, and that’s a good thing! We need to keep some value in the Premium membership as we allow Basics to buy land, and right now, the stipend value deal is the only way we can do that. Later, when we limit inventory for Basics, or provide better business reporting tools or other benefits to Premiums, then we can reduce stipends. So I agree with Ian’s argument, but I just don’t thing it’s a bad thing now.

  5. Actually, the stipend number Gene cites too low because it doesn’t count premium stipends paid to users who downgrade before the end of the month. Here’s where I got mine (less the 10% that goes to basics):
    https://osiris.lindenlab.com/reports/index.php?section_id=42&reports=53

    I get the arguement that we want to use below-market L$ sales to subsidize premiums, although I don’t really agree with it. We should probably track the value of that subsidy and apply it to the marketing budget or something.

    As for Jim’s comments – I don’t care what the residents think, and I think we should ditch stipends immediately. But accelerating the debasement of the L$ ahead of that is grossly irresponsible. Unless meeting our revenue targets is more important than keeping SL healthy? Is anyone going to make that argument?

    Once we are selling L$, the amount that we sell each month should be decided by a known algorithm or impartial entity, and *definitely* not by someone trying to meet a revenue target. It’s the only way to avoid the largest conflict of interest this side of Dick Cheney’s energy policy.

  6. We don’t exactly have to do $5000 per this statement:

    “I’d target the X% so that the $Y sale amount equals about $5000. And to the extent this proves a stable selling strategy, we’d slowly move that X% up.”

    We could set X% so the the $Y sale equals about $100, or even $50 or perhaps even $20 or less. Let’s just move forward on this in some manner. We can then come up with a metric to move or not move the X% . I’ll concede that it might be ‘attractive’ to some to set X% extremely low while we ruminate over every other little piece of the puzzle.

    But we’ve got to get over this L$ sale hurdle though and move into the X% management phase.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s