Nirmit
Sanjeev
Raju
Kartavya / Abhi
SriRam
11-09-03
(1/17)
Pricing
of our Webservices
Enclosed
find yesterday’s news-cutting which says Apple have succeeded in “selling”
(i.e. permitting online downloading) from iTunes —
10
million songs in 4 months (16 weeks).
In
my earlier notes on Napster / Apple comparison, dated May 15–16, I had
pointed out that Apple had sold
2
million songs in first 2 weeks of launch @ 99 cents / song.
So
downloads have a bit slowed down, but it is still quite impressive!
$9.9
million in 4 months.
How
did Apple manage to pull this off?
Possibly
because:
- Apple
tie-up with Recording Companies offered a legal
download — avoiding anger / enmity of recording companies.
- Apple
made them partners (by paying a certain percentage of what it
earned) rather than competitors.
Record
companies had the “content.”
Apple
had the “technology.”
A
marriage of both resulted in a win-win situation for both.
- Music-lovers
were downloading “legal” versions without feeling guilty
— a large psychological motivator for users!
- Apple
priced each song so low (99 cents) that it
removed the incentive from users’ minds to become “dishonest
thieves” engaging in pirating.
Apple
apparently succeeded in figuring out the “price” (of songs) at which pirates
would turn into respectable law-abiding citizens. Apple guessed correctly the
size of the potential market and the exact price-point at which...
it
can “swing” the market in its favour.
The
Apple episode has some useful lessons for us as far as our webservice
pricing strategy is concerned.
Keeping
our initial — and ongoing — “development costs” (salary, investment in new
office, running expenses + overheads) out for a moment, let us figure
out:
COST
OF WEBSERVER
|
Item |
Per
Year |
|
▶ Rental (for 2 servers) |
Rs.
3,20,000 |
|
▶ Depreciation (@ 50%) |
Rs.
1,25,000 |
|
▶ Capital cost of 2 servers |
Rs.
2,50,000 |
|
▶ Interest (@ 10%) on above |
Rs.
25,000 |
|
▶ Cost of UPS |
Rs.
25,000 |
|
▶ Bandwidth usage |
Rs.
1,00,000 |
|
▶ Cost of proprietary
software licenses (written off in first year itself) |
Rs.
1,00,000 |
Total
= Rs. 6,70,000 / year
We
have 24×365 hours in a year (since server is “on” 24 hours, we can schedule
extraction batches around the clock)
=
8,760 hours
So
our hourly cost becomes:
Now,
let us add “development” costs as well.
|
Item |
Cost
per year |
|
▶ Salaries |
Rs.
24 lakh / year |
|
▶ Interest on Rs. 20L
capital cost of new office (@10%) |
Rs.
2 lakh |
|
▶ Depreciation on computers
+ other hardware (@10% on Rs. 20L) |
Rs.
2 lakh |
|
▶ Other overhead expenses
(marketing, travelling) |
Rs.
5 lakh |
Total
= Rs. 33 lakh / year
So,
hourly cost (for development costs)
(B)
Hence,
Total hourly cost = A + B = 76.5 + 376.5 = Rs. 444 / hour
Let
us round off to Rs. 500 / hour.
Now
if we assume that all of these 8,760 hours the server is just doing one
thing, viz.:
EXTRACTING / CONVERTING RESUMES
and
nothing else.
And
if the server manages to extract 500 resumes per hour, then our cost
per resume extracted works out to:
So
even if we charge Rs. 2 per resume extraction, we are OK!
Only
hitch is that this would mean we must get for processing, in one year:
or
12,000 resumes each day (365 days).
Question:
Can we really manage to get that many resumes to convert daily?
If
you may recollect, we had originally thought of charging
Rs.
10 per resume extraction.
What,
if any, is the “Price / Demand” elasticity for this service?
(Simple
hand-drawn graph labeled “Demand (Lakhs)” vs “Price / resume (Rs.)” — downward
sloping line from 1→20 to 7→5.)
Will
corporates process / convert many more resumes if price is Rs. 2 instead of Rs.
10?
We
don’t know. I don’t think anyone knows!
There
are no comparable products in the market which we can benchmark for our
own pricing.
There
is even temptation that since there are no comparable products, let’s fix a high
price (since buyer/user has no means to compare, in any case) to start
with.
If
the corporate subscribers discover, through usage, that they are getting good “VALUE
FOR THEIR MONEY,”
Even
at our “high” price, the demand will pick up gradually.
On
the other hand, if we find in 3/6 months that the demand is not picking
up as targeted, then we can always stimulate the demand by dropping prices.
Nothing stops us.
Does
this argument have a familiar ring?
Same
as: most cellphone companies started out 4 years ago at Rs. 16 per minute.
Then
as more and more competitors got attracted (– because of this lucrative price
–), capacity got built up, and the prices kept falling / falling / falling to
now, maybe Rs. 0.40 per minute.
And
the bottom is still not in sight!
But
when Reliance set up base with a huge installed capacity, their goal was
very simple:
Forget
what the competitors are charging.
(I
don’t think Reliance ever used competitors’ prices as benchmark.)
Simply
declare that price at which 90 % of built-up capacity gets utilized.
Rationale
was quite simple.
When
you are in a “commodity” business,
you
are in a “process industry.”
You
are in a business of delivering
DIGITAL
SERVICE (Internet / Wireless)
—all
of your costs are fixed costs!
Hardly
any cost varies with volumes.
Your
break-even volumes are, therefore, very high.
(Graph
illustration:)
A
line chart showing Revenues and Costs vs Volume (Capacity
Utilization).
- X-axis:
Volume → (0 %, 50 %, 90 %, 100 %).
- Y-axis:
Revenue and Costs.
- Fixed
cost shown as a horizontal line.
- Variable
cost starts from 0 % and rises gradually.
- Break-even
point marked around 90 % capacity.
Caption
below:
“You
make money only after you exceed 90 % capacity usage.”
And,
from page 3/4, it is quite clear that our total costs of Rs. 40 lakhs
(approx.) per year are almost FIXED!
So,
we must ensure that our webserver is working every hour of those 8760 hours!
Not
just ready to serve but actually serving.
And
we must reach this capacity utilization within 3 months of launch.
Gradual
ramping up just won’t do! There is too much at stake.
This
(achieving 100% capacity usage) is Sanjeev’s job.
The
only way for Sanjeev to look at his job —
he
is not selling a product or a service —
he
is simply selling our webserver’s INSTALLED CAPACITY!
As
simple as that.
It
is only through 100% utilization of this HUGE investment that we
have made (in people, building, hardware, software, etc.) that we can recover
our fixed costs — which is all of our costs.
Now,
at this stage, it may be legitimate to ask,
If
our hourly total cost of server is
then
why not simply “sell” this processing capacity
@
Rs. 1000 / hour?
—
irrespective of what subscriber uses it for,
which
applications he runs, what he uploads/downloads.
But
if we did that, in a year, we would earn only
i.e.
approximately twice our fixed costs.
That
seems ridiculously / absurdly low!
We
might have succeeded in filling up the capacity (8760 hours) but priced
ourselves so low that in a whole year we end up earning Rs. 80 L!
Can
we think of pricing / selling one “server-hour” for, say, Rs. 5000 / hour?
That
would help us earn
That
looks more reasonable / acceptable / justifiable (from our point of view).
But
the question is —
Who
will buy a webserver processing hour at Rs. 5000 per hour, even with
GunMine / GunSearch applications thrown in?
If
Microsoft were to offer even their best software at Rs. 5000 per hour (i.e. Rs.
40,000 per 8-hour shift), how many corporates would buy how many such
processing-hours?
I
doubt if any would.
If
we were to make such an “hourly-charge” proposal to our subscribers, we would
kill the business even before it’s born. People would simply laugh at us.
So,
we have to stick to ‘mouse-click’ based Tariff Structure.
In
this structure, each tariff amount looks so small / so insignificant that you
literally
“ignore” it!
When
you see a “Tariff chart,” your mind does not start multiplying and adding!
It
is quite like an “À la carte” menu containing 40 food items, each priced at Rs.
40 / 50 / 20 / 10 etc.
vs.
A
“lunch / dinner Thali” priced at Rs. 400 (even with unlimited consumption!).
The
high figure puts you off.
Very
often, you also think that in any case, you will be unable to consume
everything that’s in the Thali — then why go for it?
Why
not pay for only what you want to consume — even if it adds up to more than Rs.
400?
In
pay-per-use, you are able to establish a “correspondence,” a
one-to-one relation between the price and the value.
In
a “Thali” or “Per-hour” scenario, you are not sure whether you’ll manage
…to
derive the value for the money paid.
This
is why cellphone tariffs are also “pay-per-use.”
And
to entice you to use more & more, service providers are daily coming out
with ever-newer “uses” and additional “tariffs” for each use.
It’s
the same principle of pay-per-use in:
- ▶ Electricity →
Units consumed
- ▶ Gas → Cubic
metres (c. m.)
- ▶ Water → Cubic
metres
- ▶ Travel → Km
travelled × class of travel
- ▶ Freight →
Volume used (cubic feet) or weight loaded (tons) (again per km taken)
“Per
unit” of consumption/usage, the amounts look insignificant / small.
And
this attracts both small users & big users.
They
know that their damage / risk / commitment / cash outflow etc. is
something they themselves are going to control.
Anytime,
they can speed up or slow down.
They
are in control at all times.
And
especially, if you are not “watching” the meter continuously, you are not too
much perturbed.
A
classic example is that of a self-service supermarket.
Goods
are arranged very attractively, tempting you to pick up & add to the
“shopping cart.”
Although
you do look up the price when picking up, you are not carrying a calculator
& “adding up” as you move along the aisles — happily picking up even those
things which you have no immediate need for!
Quite
often, it is only as you are passing by an especially attractive display that
you realize you need that item or that it would make a nice addition to
your bedroom / living room / kitchen / wardrobe etc.
It
was not even on your shopping list when you left home!
But
now you want it just because it is there in front of your eyes!
Shopping
malls first CREATE a need — then fulfill it!
Quite
often, it is a question of “pride of possession”
and
has nothing to do with any real or any “perceived” need!
Digressing
a bit,
Once
its capacity utilization (of its network) exceeds 80%, Reliance — &
all others — will slowly raise their prices once customers have got “hooked”
(no portability of numbers!).
And
their price rises will be quite small but frequent (2–3 times / year) so that
users won’t even notice that their monthly bills are slowly creeping up, month
after month!
After
all, as a consumer, you know that costs/prices of all goods & services are
rising by 8–10% each year, so you don’t get agitated. A long annual price-rise
is within these limits — your systems are able to absorb/adjust such increases.
This
equally applies to individuals as to corporates.
Hence,
our pricing strategy shall consist of:
▶ A wide variety of
“products” (buttons)
▶ Each priced low /
insignificant in order to make it readily acceptable to users without batting
an eye.
▶ Continuously watching the
“response” (daily no. of transactions for each type of transaction) to
forecast:
- which
transactions (products) are popular / in great demand & with whom,
- rate
of growth of each.
This
must be watched “subscriber-wise” & “overall.”
▶ Based on such daily watch,
periodically increase the “tariff” just so slightly that customer does not even
feel the difference (except, maybe, on a year-to-year basis).
▶ Do not show to a
customer the METERS (transaction-wise or overall) — except them being
logged by ADMIN only.
▶ (Not part of pricing)
— Arrange products (buttons) very invitingly / very suggestively,
tempting him to click! (Remember,
“Need”
must not be a criteria for clicking!
As
Mallory said,
“I
climb Mount Everest because it is there!”
▶ Keep adding new buttons
(new services)
and
announce these additions on homepage (which is our “storefront”).
▶ Create a buzz about:
“You
will be left out of a GREAT! GREAT experience if you do not subscribe.”
(“Keep
up with the Joneses” syndrome.)
“You
are in great company when you do.”
No comments:
Post a Comment