They *will* pay--living proof [long]

From: Tom RC Campbell <admin_at_esnipe.com>
Date: Wed 13 Jun 2001 10:15:24 -0500

I own and manage a site called eSnipe, which places
bids on eBay the last few seconds of the auction.
Please forgive the seeming plug, but a description of
the site's specifics are essential to this article. I
have been following the "Will They Pay" threads with
more than a little interest. I bought eSnipe, then a
free service, last December assuming that:

1. People were willing to pay for a site that performed
a useful service.
2. The ad market would tank.
3. I would conceivably have to rewrite the site itself
due to performance problems just before the sale was
announced, leaving me with little more than a brand
and 24,000 registered users.

All predictions came true. Shortly after I bought eSnipe,
which had been run by a single intrepid hobbyist, I
hired an expensive but brilliant net guy to rewrite
the site from scratch, which he did--while it was
running. eSnipe is particularly expensive to run
because of its real-time considerations. If eBay lags
for a few minutes or goes down for a few hours, they
don't really care that much. But if my site slows for
even a few seconds, the damage can be catastrophic. So
we have dedicated servers with 24 hr staffing in 3
different parts of the country, with 3 different web
host providers, a fourth server for mail, and a fifth
for DNS, to provide maximum security and reliability in
case a server fails, along with incremental database
backups every few minutes and failover code to move
loads to different servers if one fails. Need I say
this makes even updating a news item on the home page
complicated? Because now it has to go on *3* servers,
so we also have our own publishing system too.

Other than the net guy and server staff, there's just
me, doing support, marketing, etc. More on marketing
later. I handle something like 40 support requests a day,
the vast majority being due to lack of knowledge on the
users' part about how eBay works (but I still investigate
each one, just in case). The site had 60,000+ registered
users and placed 8,000 bids/day as of May 30,
representing $100 million in won bids, plus another $100
million in bids that were too low. We announced that we
were going to charge 1% of the winning fee as of
Memorial day, with wins under $25 free (that's 60% of
the bids we place), and a limit of $10 on any one fee.
There's no charge if you don't win the auction. The pay
system announcement was of concern for two reasons.
One is that we employ a homegrown micropayment system
called BidPoints, primarily because the average fee for
an auction win we charge for is 40 cents, and that's
about what credit card processing costs, so we require
that you pay for your auction wins using BidPoints. The
second is that we originally announced you'd have to pay
in advance for BidPoints--my feeling was that with our
extremely low margins, we couldn't afford deadbeats.

Of approximiately 275 responses I got regarding the pay
system during the 6 weeks between the announcement and
its debut, the breakdown was:

    Yes - 110
    No - 38
    Confused - 30
    Bill me, don't make me pay in advance - 50
    Can't use credit card - 15
    Questions - 30

The "Bill me" mail had to be taken seriously and required
a two-week delay in the pay service debut because it
represented a change in the whole payment architecture.
This request was from heavy users who just didn't want
to be bothered; they didn't want to be forced to buy
up front, and they didn't like the service fee I charged
if you forgot to pay in advance. Very little of the
"bill me" mail was from the kind of flakes who just like
to point out rounding errors in my examples and misspelled
words in the FAQ.

As far as marketing. Technically eSnipe isn't marketed at
all. We've never paid for an ad, nor have we paid for
search engine placement. Indeed, the most common sentiment
in fan letters is that they love our site, but they
don't plan to tell anyone about it! I have always planned
to buy keywords in search engines, pay for a Yahoo listing,
etc., and I hope to start that soon. Still, we get well
over 100 new registered users every day, all through word
of mouth. But the main element of our marketing has been
making this a reliable, easy-to-use site and providing
customer support. The previous owner had a full-time job
and stopped providing support for eSnipe. I on the other
hand saw support as a soft-sell marketing opportunity, and
I also wanted to get an idea how mch support would cost,
should I need to hire someone to do it for me. Clearly I
need to do the more traditional stuff, but until recently
I wasn't sure our site could handle a heavier influx of
new users than we already get.

My net guy, whom I respect greatly from a technical
standpoint, was sure eSnipe would go down in flames when I
started to charge. He has experience on several other,
much bigger websites, and was certain users wouldn't pay,
or at least pay enough to keep up the payroll. He suggested
that I start another site under a different name and charge
on that one, while keeping eSnipe free. My response was
that eSnipe, already the "market leader" (if nonpaying
users can be considered a market) in sniping services, was
a strong brand name that users trusted, and that I had no
intention of flushing that hard-earned branding down the
toilet. He then suggested I charge a minimal amount for
every snipe instead of a percentage of wins. This was
because of the incredibly difficult conceptual and
architectural problems associated with charging a percentage
and tracking down wins (we have to do this all by scraping
eBay's HTML pages, since they won't, understandly, let us
use their presumably-more-reliable developer's API). But
I felt there would be a fundamental change in our
customers' habits if we charged for every snipe, winning
or not, even if it was a small amount (he suggested 10
cents), whereas they saw real value in a win. And by
charging a percentage we had a pay-as-you-go system--
cheaper than a subscription for casual users, yet not
overly expensive for big users because of our $10 cap and
a discount on bulk purchases of BidPoints. Our
disagreement was fundamental: I felt intuitively that our
users would feel that eSnipe was worth paying for and
would largely remain loyal. He felt that the lopsided
response to the pay system was simply because most users
would quietly leave the second we started charging,
without ever letting me know why.

Here's what I expected to happen when we flipped the switch
on Memorial day.

* Fiscally, I planned for an 80% drop in usage. Yet I also
had to assume that our current signup rate of 150 users/day
would continue from a logistical point of view, because what
if the "worst" happened and our traffic stayed where it was?
It would be a disaster if by some miracle traffic stayed high
while we charged but we hadn't allocated enough server
resources to handle the business. So I had to both assume
we'd lose a lot of business and that our expenses would
remain as they had been.

* Privately, I expected a 40% drop in usage. We all "know"
that you can't charge for a website, don't we? Still, I
thought 60% of our users would stay on, despite the fact that
there are at least 3 other free sniping sites.

* My wife, who does no work on the website but who discusses
strategy with me, expected a 20% drop in usage.

* Because we don't charge for bids under $25, I expected the
proportion of those bids to increase dramatically.

* I expected buying activity to be slow, and slowly build
for the next two weeks. To encourage buying I instituted
temporary 20% discounts at the $40 level, and 25% at higher
levels. Since $40 buys you $5,000 in fees for auction wins,
I didn't really expect much activity. Also, we purposely did
not state whether you had to buy BidPoints up front, and we
also stated that you did not *have* to buy BidPoints until
2 weeks after your next login or July 1, whichever came first.

On Memorial Day, when the pay system was supposed to go into
effect, disaster struck. The first day we started charging
for our site, receipts were 30 times what I expected (a mere
3 times what my wife had expected). Business slowed a bit the
next day, but not by much. They have leveled out, after 10
days, to about 10 times what I had originally expected to
earn per day after my first year. I'm sure that receipts will
go down somewhat, but it's a big success. The "disaster" part
is that we haven't fully automated the pay system, so it takes
a long time to reconcile purchases with info from Bank of
America and PayPal, our merchant account processors. Of course
it's a nice problem to have, but it has increased my support
(tough life). Meanwhile we're frantically working to complete
payment automation.

Site traffic decreased about 10%, not 40% or 80%. Bidding
habits did not change at all--the high-ticket bidders are still
bidding like mad. Average purchase was $20, and .5% of our
unique logins have been buying each day.

Lessons I learned:

* People will pay for a site they value. Part of it is the
mystical appeal of online auctions, which I still don't
understand fully. But there are sites for whose content I
cheerfully pay ("Consumer Reports" and a guitar-related site,
for example). I was willing to believe that I was unusual,
but I think not.

* I think part of the success is customer support. For the
last 5 months I have tried to resolve every support incident
and have come pretty close. Most of it can be handled by about
a dozen form letters I use, but they're well-written and often
get a minor tweak for each customer. I realized that at about
300 support incidents per week, that's 6,000 since the first of
the year. I've interacted with a good 20% of my active customer
base. So they know me.

* There's no free lunch. My site costs me $30,000/month to
run. If you're a programmer, you'd think that what we do has
what venture capitalists call a "low barrier to entry"--that
is, it's so easy, anyone can do it. NOT! It's easy to write
the first 80% of a sniping service. But eBay constantly
changes its web pages, they add new auction formats, they
acquire new companies, they sometimes screw up the operation
of their site, and so on. We need a full-time programmer just
to keep up with these problems, and a full-time network
administrator to react to eBay outages, eBay servers down, etc.

It looks like our site will be meeting its payroll (including
my salary, which I not taken until now) and then some. On the
one hand, that's great, right? But on the other hand, our
staffing is extremely lean. 4 full-timers do everything. If
you heard that a site like mine was being funded by venture
capitalists, do you think they'd be able to do it so cheap?

* Automation is good. The 4 full-timers can easily handle 4
times our current business. By the time we need a 5th employee,
she'll already have been paid for.

* Listen to my wife. No, that's not just a glib joke, although
it works that way too. She wanted to go pay much earlier. I
wanted to go later because not all of our automated support
systems were perfect. By my unrealistically bad worst-case
estimates, we wouldn't turn a profit for a year. So there was
no rush to go pay. My wife figured we'd be profitable from
Day 1, so she understood that we were losing money every day
we didn't charge for the site. The point is, I *was* funding
and running the site properly. It was not reasonable to
expect the worst case because we weren't wasting money, we
were providing good customer service, we had invested huge
sums in designing and engineering the site, and we were
consistently getting 1,000 new registered users a week. It
was okay for me to assume a drop in business, but not 80%.

So it *can* be done. You can charge for your site, if it's
worth paying for.

Sorry for the length of this letter, but there were a lot
of things I wanted to get across to this group that has
given so much to me over the past few months.

Tom RC Campbell



Received on Wed Jun 13 2001 - 10:15:24 CDT