They promised us an "Infinite Jukebox" -- but we never expected the price to be infinite
The early days of the Internet promised an "infinite," "
celestial jukebox," with instant access to all the content in the world. But instead of heaven, we are now facing "
subscription hell." Yes, we can now enjoy nearly infinite access -- but the price also seems to be approaching the infinite.
What we have here is not a failure of technology but a failure of business model
innovation.
The future of subscriptions is to make them risk-free to the consumer. For digital services, the provider risks little except the opportunity to take money in exchange for no value. That will be less and less tolerated.
===================================================================
Your thoughts? This is still in formative stages, and feedback is invited.*
===================================================================
One issue is that subscriptions are all about relationships, and that is more true than ever in our digital world. Our current relationships are dysfunctional because
businesses
make consumers take on pricing risk for no good reason. Consumers see the risks
they face, realize that makes little sense for digital services, dislike that, and dislike businesses for demanding that.
The compounding issue is that the inherent abundance of replicating digital services makes consumers even less willing to accept pricing risk.
Providers seem to think current models are the only way to do
business – if they think about it at all. (Some prefer not to think about it, and love the value proposition of "autopay forever," hoping you forget that they are sucking money out of your wallet every month.) Even those with better intent and more desire to innovate are stuck in the scarcity-based economic mind-set of real goods, and have not really understood the value-creation power of the new
economics of abundant digital services.
We are still mired in your father's subscription models -- models for mailing pre-defined assemblages of print and squeezing pre-defined sets of TV channels into an analog cable:
- Business know that consumers like simplicity.
- They also know that consumers hate surprises.
- So subscriptions are made simple: unlimited, flat-rate, all you can eat (AYCE).
- But AYCE distorts incentives -- it overcharges light users and undercharges heavy users. It limits risk at the high end, but not at the low end or the middle.
Businesses know
there are problems here.
- They have difficulty acquiring customers, and so offer introductory discounts (jam yesterday)
- They have difficulty retaining customers, and so offer retention discounts, but only after you try to cancel (jam tomorrow)
- But there is rarely any discount in normal times (never jam today).
Seeing the problems with AYCE subscriptions, some turn to another unrealized dream of the Internet -- like an old style jukebox, our infinite jukebox should take nickels -- so-called "micropayments."
- But micropayments just change the pattern of risk: how many nickels will I need?
- This reduces risk at the low end, but not at the mid end -- and dramatically raises risk at the high end.
We have limited forms of micropayments for decades, in the form of pay per item (PPI) or pay per view (PPV). But, run up enough micropayments and those digital microbucks add up to real kilobucks. That is a fatal problem, even if we can make taking the micropayments totally frictionless. (Many grasp at new hope from cryptocurrencies and blockchains, but, much as they may reduce friction, they do not solve the problem of risk.)
The problem remains: consumers hate risk! Both classes of current models force significant and unnecessary pricing risk onto the consumer.
As outlined here, we can easily do much better, for most kinds of subscriptions to digital content or services -- not only for the consumer, but for the publishers (and platforms that serve them).
By getting smarter about harnessing the abundance of digital, we can have our cake and eat it too: we can reduce consumer risk AND we can motivate consumers to sustain those who create the content and services that they desire. (See "extensions to sustain creation" below.)
What is a risk-free subscription?
Here is the basic idea of a "risk-free" (or "no-risk") subscription. Compare it to a conventional unlimited subscription that costs, say, $5/month. (The "update" section below also considers how this compares to services that are currently free and wish to shift to a paid model, or to enhance a simple voluntary payment model.)
Let's design a risk-free subscription that costs $0 to $7 per month depending on how much you use. Let's design a volume discount that varies -- to work much like micropayments for low usage -- and much like an unlimited subscription at high usage -- with graceful blending in the middle.
- Get "run of the house" access to whatever items you want
- If your usage for the month is zero, your bill is $0
- As your usage for the month grows, your bill grows, but with declining cost per item. Your bill will go from $0 to $7, depending on how many items (and how many of them are premium items).
- To avoid the risk of bill-shock when you used more than you intended, you never have to pay more than $7.
This is a simplistic example, and the price ($7 versus $5) for this added flexibility may actually be reduced over time. If many more people subscribe (because they have lower risk), total revenues will grow and the ARPU (Average Revenue Per User) target can be reduced (to attract still more subs), so the unlimited cap might shift to $5, or even lower.
We can improve on this (as explained further below):
- add nuance to our usage metrics to move us closer to a value-based metric that understands that some clicks are more valuable than others
- layer on options to more fully support the ongoing investment of publishers and creators.
(This risk-free subscription is a generalization of a model I first suggested in 2015 --“
Post-Bundling – Packaging Better TV/Video Value Propositions with 20-20 Hindsight” -- and have since discussed with major TV providers. That provides added detail on the use-case for TV/video bundles.)
[Update] Note that while the discussion in the original post was about risk-free models that ramp up
based on items accessed or consumed, the same principles can be applied to work
based on time, not items. Some possible advantages of that are outlined in the update section on time-based models at the end.
Of course such a subscription is not absolutely risk-free, but it is much closer -- and yes, there are some levels of risk to the provider -- both of which are discussed below. But first, a closer look at consumer risk.
Consumer risk in an unlimited subscription
Think about the consumer's issues when they decide whether to subscribe, ...as they continue, and ...if they consider cancelling:
- Will I use enough to justify the monthly price -- now, in the past, and going forward? Am I using the service often enough?
Am I happy with my interest level in the selections offered?
Am I satisfied with the quality of the items I consume?
Do I just skim many items, or quit part way through?
Do I get the desired value (or enjoyment) from the items?
- Which premium channels should I buy access to?
How would I know in advance?
Did I watch enough items on the premium channels I chose and paid extra for?
Was I happy with the premium channel items that I did consume?
Did I regret that I could not watch programs on premium channels I did not subscribe to?
- Is this subscription one that deserves to be in the "portfolio" of sources I pay for unlimited access to (given all the content sources of this kind that I want)?
Did I find this month that I wanted other services I did not subscribe to?
Can I afford to add still more subscriptions?
Is this a subscription I should drop, so I can afford something else?
- How can I predict any of this reliably?
Do I know what will be offered in coming months?
Do I know what alternatives will draw my attention elsewhere?
Will I be paying for periods where I am on vacation or too busy?
The problem is that most digital consumer services offer constantly changing collections of experience goods. Especially for content services, we have only limited ability to predict what value will be offered, what items we will actually choose, and what value we will realize. That is highly unpredictable, except in hindsight.
Subscription providers seem to ignore this. They focus on customer acquisition and customer retention (and its converse, churn), but how many of them consider the dynamic value propositions of value/risk to each individual consumer? They optimize for CLV, the Customer Lifetime Value
to them, but not for
VLV, their Vendor Lifetime Value to the customer. How many businesses really think about how they justify their share of the consumer's wallet?
As more and more content of all kinds goes behind flat-rate subscription paywalls, how many subscription are
simply unaffordable to many consumers who might gladly pay a profitable amount for occasional access? How many services offer discounts only for new customers, or those threatening to cancel? (This reflects a natural
risk discount -- if the price is set in advance, a consumer's willingness to pay must factor in a discount to adjust for their risk of disappointment.) What about those who would be continuing customers at modest but still profitable levels? A few top publications like the New York Times are making money with subscriptions paywalls --
but only about 3% of their readers subscribe! -- and most news publishers do much worse. What a waste to both consumers and businesses! Surely we can do better!
The game most subscription providers play now, is one that charges at flat rates that work
for their best customers, but that leave more moderate customers on the ragged edge of saying no. And the vast majority of those who might pay for moderate amounts of content
do not subscribe at all. Some providers are even more cynical and customer-hostile, hoping you will take a trial and forget you are paying $5 a month, and then making you jump through hoops when you realize you no longer want to.
Consumer risk in micropayments (pay per item)
Why do consumers hate micropayments? -- even if they are frictionless? Because consumers hate unpleasant surprises.
- What if I run up a huge bill?
What if I get hooked on a binge?
What if my kid goes overboard?
- Will I be sorry I paid per item instead of getting an unlimited subscription?
- What if I select items but find them disappointing?
- What if I like to skim, and so access many items but get little value from each?
These problems are inherent in micropayment models that do not have significant volume discounts or other value-adjustment provisions. Pay per view movies have a profitable niche, but viewing more than a few gets very costly. News services like Blendle have been even less successful -- they offer single articles, but at 25-49 cents each, the bill rises quickly.
The psychological distress of the ticking meter has been well established. Think of telephone minutes, cellular data megabytes, and the old days of online minutes on AOL. Knowing the billing clock is ticking makes it very hard to enjoy using a service. We are always worried: "what shock will I face when I see the bill?"
20/20 hindsight and post-pricing
My work on
FairPay highlights the difficulty of setting prices before the experience, why that is an issue of risk, and how "post-pricing" can avoid that problem. The value of experience goods is best known with 20/20 hindsight. Consumers are much happier paying for the value they get after they know what the value actually is. The classic Our Gang "
Pay as You Exit" story illustrates the power of that.
Provider risk and profit
Back to my opening statement, "the provider risks little except the opportunity to take money in exchange for no value." Providers will, of course be quick to argue that they do face risk, but to what extent? Since the unit costs of access to existing content and most other digital services is negligible, the risks are not the marginal costs of service, but the usual subscription issues that drive CLV -- CAC (customer acquisition cost) and retention/churn -- and the risk of just not having enough subscribers.
The deeper provider risk issue relates to the predictability of cash-flow -- whether they can expect to fund their content creation and marketing expenditures going forward.
- Compared to micropayment/PPI models which are already totally dependent on usage, the risk-free model should not worsen predictability -- and might improve repeat activity enough to make predictability much better. Many businesses are hit-based, and deal with it, and all but the smallest publishers can spread that risk.
- Compared to flat-rate subscriptions, the obvious concern is that the steady stream of monthly payments from each customer might become much less steady. However, the law of large numbers (many customers) will tend to smooth that in aggregate. Also, if the risk-free offering is designed well, there is reason to expect that reduced CAC and churn will dramatically increase the number of subscribers, so that overall revenue and net profit will be much higher, even if it is more variable.
Cynical providers be very reluctant to shift from the "get them on autopay and hope they never think about it again" gravy train -- but isn't that really very thin gruel?
Tuning the model
That is the basic idea, and the core of the case for it. It will take good design, testing, and refinement to prove it out and get it to work well. That can start with limited, low-risk tests. There is reason to expect that to validate some promising sectors and customer segments, so it can grow from there.
The rest of this draft explores some ideas on how to build on this strategy, by further reducing consumer risk and adding more
value-based metrics of usage -- and by keeping the impact on provider risk manageable.
Value-based pricing is increasingly viewed as best-practice in B2B -- we need to be more creative about applying that for B2C.
The discount curve
A key feature of the risk-free subscription is that it depends on usage, but adds a volume discount. Designing the discount curve that gets built into the price schedule will be critical to making it behave in a way that can make consumers comfortable. Consumers want simplicity, so the trick may be not to expect consumers to look closely to understand whatever tiered or continuous schedule of rates is used, but to simply give some examples of what to expect at representative usage levels. As long as customers have a sense that the curve is reasonable, and that they can see their detailed accounting for any month if they want to, that may be enough -- as long as the cap on total rate is not too high, and they don't reach it too quickly. (Of course one or more levels of premium pricing might be reflected in this schedule as well.)
Extension: The money-back guaranty, and the skim discount
One thing Blendle, the news micropayment aggregator,
did well was to offer
an unconditional refund button on each article viewed. That is a good start, but too all-or-nothing. It may be much better to let users specify a percentage refund they want, so that they can ask for a
partial "satisfaction discount" when they are disappointed, without shying away because they feel a full refund would be unfair if they did get some value.
(Note that a quality guaranty can
increase willingness to pay, thus offsetting the cost of the guaranty. Now consumers unconsciously build in a risk discount that discounts for the risk that they will regret their purchase. The guaranty can eliminate the need for customers to discount for that risk.)
Related, is
the skim discount.
Instrumentation increasingly makes it practical to determine the time spent consuming items, and what portions are consumed -- why not discount the unit price if the time spent is clearly short, or the item is clearly not finished? This kind of tracking also makes it possible to confirm that subscribers are being honest about claims of dissatisfaction, and limiting refund privileges for those who go overboard.
Extension: value-based usage metrics
Advertising-based revenue models lead to click-bait, and there are valid concerns that usage-based revenue models can create the same kind of harmful incentives. A simplistic usage metric such as number of items accessed, may well create similarly misaligned incentives for quantity without quality. But extensions like the satisfaction discounts and skim discounts above, will shift this from a simple count, toward a more nuanced value-based metric.
Further extensions can add more sophisticated value metrics (and the bonuses of the next section) to make this model more reflective of the true value of the experience to the consumer. Such metrics may factor in time spent with an item (dwell time), how full a portion of it is consumed (aborts and sampling), is it re-accessed, does it lead to further actions (outcomes), is it shared, etc. Of course most users will not want to dig into this complexity, but a simple "relative value/intensity of use" metric for each item could be reduced to an average and included in their statement. That is likely to be accepted as long as it seems reasonable.
Extensions to sustain creation:
A publisher-sustaining bonus, and a creator-sustaining bonus
The real challenge in sustaining digital services, especially content services, is that we are only beginning to realize that we must have a new
social contract.
We must pay to sustain the supply of future content, which is costly,
not to access current content, which costs almost nothing. A risk-free subscription can make this transparent and discretionary:
- At the end of the period (along with the statement that reports on usage, and what the "risk-free" price came to), invite a voluntary bonus to sustain the publisher.
- Remind the subscriber what they accessed and what they apparently got the most value from.
- Invite them to add a bonus payment, to reward the publisher, to better enable them to continue to supply more like that.
- Also, invite them to make this a recurring bonus (that can be cancelled at any time), so the publisher has more certainty of continuing revenue.
This can work for single publisher subscriptions, and for aggregations. Aggregators can suggest that bonuses be contributed for each publisher the consumer patronizes heavily (as well as a bonus for their own curation services).
A similar bonus can be offered to reward and sustain creators/artists -- the authors, musicians, filmmakers, gamemakers, or podcasters that each customer patronizes most heavily. Report the top candidates each month, and encourage a voluntary sustaining-bonus contribution that goes directly to them -- one-time or continuing. This might substantially increase consumer willingness to pay, and might generate significant
benefits down the value chain, to enable digital services to create value far more sustainably.
This would work as a new kind of hybrid model, adding a component much like recurring crowdfunding (as supported by Patreon and similar recurring variations of Kickstarter and Indiegogo) into the mainstream of subscription businesses. Of course these bonuses need not be entirely voluntary -- there could be some required minimum "sustaining fee" -- or some premium-level sweetener could be added that requires a minimum fee.
A low-risk step for publishers in the direction of FairPay
Notice how this risk-free subscription becomes a way to edge toward FairPay while limiting risk to the publisher. The simple no-risk subscription outlined here uses the 20/20 hindsight of post-pricing to largely eliminate the consumer risk in conventional pre-priced subscriptions (and micropayments). It does that in way that
keeps the provider in full control of the price schedule. FairPay goes farther to reduce consumer risk, in a more unconventional way, by
adding customer participation in setting the price. Powerful as that
promises to be, it is understandable that many providers are hesitant to give up that control.
The extension features outlined above gives the customer limited power to effectively adjust the price. They can adjust
downward with the guaranty and
upward with the bonuses. That moves incrementally toward FairPay, with a basic level of participation in a portion of the pricing.
It may be hard for publishers to make the case that they should be able to "take money in exchange for no value" -- but they do have a legitimate case that if they must invest to provide the value that consumers want in the future.
A consumer who values the service has some obligation to sustain that investment.
- Think of the base risk-free subscription as the way to maximize market reach, and ensure a base level of compensation commensurate with usage. (A component of price that is controlled by the provider.)
- Think of the sustaining bonus as the way to nudge consumers to sustain the ongoing creation of services they value. (A component of price that is controlled by the customer, within limits set by the provider.)
Think about where to start with this, test it, learn how to manage it, and move toward a solution that serves more consumers and generates more profit for providers.
And, with this kind of win-win model, we can more sensibly sort out a balance between publisher-specific subscriptions and aggregated services (as
Apple has given added prominence to) -- to find a harmonious mix that is good for consumers, publishers, and aggregated distribution services across a full spectrum of dynamically varying usage levels.
Why not give it a try?
---
Update: A "pay-ramp" not a "pay-wall!"
Think of the risk-free subscription as a "pay-ramp" -- a gentle incline that presents an almost imperceptible barrier to new and light users that permits ongoing sampling at modest cost as long as desired. As users gain the usage habit, the cost ramps up slowly, and even bingeing is at low risk.
Compare this to the sudden, hard barrier of a "pay wall" -- Even so-called "soft" or "metered" paywalls are really hard, shutting the door once you exceed the quota of free items. Instead of a hard barrier, the risk-free pay-ramp is more like a gentle speed-bump -- it does not stop users, just slows them momentarily.
Compare this also to an introductory subscription discount that has a somewhat lower (but still sharp) barrier, and then a further sharp hurdle (even if largely hidden, but overhanging) when the discount period ends.
Update: Time-based risk-free subscriptions -- with rollovers
Some services offer time-limited access, "pay as you go" models that are a form of micropayment that is not item-based but time-based. These can include "day passes" or variants for some number of minutes, hours, days, or weeks.
I have not been keen on the "ticking meter" aspect of these (the same problem as
other micropayment models), but in speaking with a SaaS provider of such services I learned that some services are getting much better results with these than with item-based models. That led to discussion of a time-based risk-free model.
- Current time-based models require the user to choose in advance whether they want an hour, a day, a week, or more. Interestingly, users often favor the shorter options because of fear that they will not use the longer options (even if it offers a hefty discount) -- lose-lose on both sides.
- So of course I proposed applying a post-pricing variation. Why not provide automatic rollover? Let the user start with the 15-minute or 1-hour option, but if they want longer access, don't go back to zero, roll into the longer, more discounted, option and credit the short-term fee against the longer-term fee so there is no risk.
- This eliminates all but very small price-risk hurdles, and provides full optionality.
- Just provide a simple notice to the user when the current period has expired, stating that continuing will trigger the rollover.
- Being time-based also has the advantage that it can seamlessly roll right into a full, auto-renewing subscription. On each monthly roll-over, just give the user the option to make it auto-renew.
- The UX need not even explain all of this up front, it can do it in incremental bites as rollover points are reached.
Update: From free to paid (whether compulsory or voluntary)
The discussion above was largely framed as an alternative to paid subscriptions (or pay per item), this risk-free model is attractive in comparison to other conventional models as well.
It is very relevant to services that are now free, but that want sustaining revenue.
- As an alternative to adding a conventional paywall, the no-risk value proposition softens the blow. Users can more easily be converted to paying customers if they know they do not risk paying for value they may not get. The hurdle becomes vanishingly low. Try it and keep usage low, until you see how it prices out. Instead of $5/month every month, it may be $0 or $1 or $2. If the discount curve is not right, usage will be driven down, but not to zero. That buys time to tune the discount curve so that most free users can be converted to paying customers. Make it easy to see the current total (and to get item refunds), so reluctant customers need not fear any billing surprise.
- For mission-oriented providers who want all payments to be voluntary, the sustaining bonus component does the job. Keep the subscription price at zero, but suggest sustaining bonuses to the service provider, and to selected creators of items that are accessed. For example, a public service news offering could know its customers and gently nudge them based on their usage patterns, and any other value data it can apply. Patrons can be encouraged to set up recurring bonus contributions, knowing they can be stopped or adjusted any time they like. They can be nudged to boost their bonus contributions whenever their observed value consumption increases.
- For member perks, this brings a risk-free way to charge for premium membership tiers. Many services (whether for-profit or non-profit) find it challenging to add paid premium tiers because the value propositions are especially lumpy, making fixed contributions especially high-risk to potential patrons. Risk-free charging solves that problem.
Update: How this can enhance and complement advertising revenue
It should be noted that this enhances and complements ad-based revenue models in two ways:
- Paywalls conflict with ad revenue because they dramatically reduce views. Risk-free subscriptions reduce that effect because they reduce the subscription hurdle. They are more like an pay-ramp or pay-bump than a pay-wall. Paywalls reduce reach, and thus page views. So even if direct ARPU decreases, when ads are factored in, ARPU can be expected to increase.
- The win-win nature of risk-free subscriptions can be complemented by a similarly win-win model for advertising. That is to apply a "reverse meter" that gives users credit for the attention and data they contribute when viewing ads. My post, Reverse the Biz Model!, explains how this can re-align incentives to make advertising more valuable for users, advertisers, and publishers/platforms.
Together, these strategies can have a compounding effect in raising revenue.
Update: Risk-free aggregation as the savior of long-tail providers
"Subscription hell" and "subscription fatigue" are especially limiting for long-tail providers. It is hard enough for dominant providers to attract subscribers, but even harder for smaller providers of more niche content. They are more value-challenged in attracting subscribers and more hungry for them. Potential subscribers face a lumpy value proposition (an expectation of fewer items of interest per month than for a dominant supplier such as Disney or HBO or NY Times) that makes flat-rate AYCE a very high-risk. An all you can eat buffet has low appeal at any viable flat-rate price when there is not much you want to eat.
Risk-free models can be especially valuable to them, and an aggregator who supports such models can help them reach beyond the small core of customers who would cross the hurdle of a flat-rate subscription (and not quickly churn away). This builds on my closing paragraph above: "this kind of win-win model...can...find a harmonious mix that is good for consumers, publishers, and aggregated distribution services across a full spectrum of dynamically varying usage levels."
Detailed discussion of how risk-free aggregation can work in the context of TV/video subscriptions (equally applicable to other content types) is in my earlier post, “
Post-Bundling – Packaging Better TV/Video Value Propositions with 20-20 Hindsight.”
Update: Paywalls soft and hard versus risk-free payramps
NiemanLab reports that soft paywalls will get weaker with the change in Chrome that will prevent detection of incognito mode. The suggest this "could encourage more publishers to go all in on a hard paywall, in which you can’t read a single article without first registering."
As I commented, "Maybe the solution for publishers is to shift from the zero-sum thinking of warfare
against their readers to the win-win thinking of co-creating value
with them."
Update 9/30/19: Some corroboration from Zuora
An interchange with Jessica Lessin, triggered by an article in
The Information reminded me of some important corroborating data from Zuora, a leader in services to SaaS businesses that analyzes data on over 900 SaaS companies (both B2B and B2C). From their chief data scientist, in "
What Goldilocks Can Teach Us...:"
“Giving subscribers too many options is overwhelming. Giving them too few options is unattractive to customers who demand choices and value. But there’s a just right sweet spot based on understanding how your customer values your offering that leads them over time to sign up for more. A case in point lies in how you charge subscribers for usage with a pay-as-you-go component to your billing plan: without usage billing, subscribers may feel that a one-size-fits-all plan is charging them for more product than they use. But if you charge primarily based on usage, subscribers feel like you're looking over their shoulder and charging them for everything they do.”
“Our research shows that churn is lower and companies grow faster when there is a usage component to pricing: these companies experience 6% lower churn and 8% faster annual growth. But faster growth happens when the usage component is less than 50% of the bill: an additional 4% annual growth compared to companies where usage is the main mode of billing.”
(With regard to the last point, that "faster growth happens when the usage component is less than 50% of the bill," it seems unclear to what extent that really applies to B2C markets. But to the extent it does, that would suggest using the risk-free "pay ramp" as I outline it here, but perhaps with a minimum "floor" price per month even when there is no direct "usage." This gets to the issue of more sophisticated value metrics beside usage, and might be framed as covering the value and cost of "continuity" services, such as curation, alerts, newsletters, and even optionality. But such a floor price should not make the perceived risk of overpaying too high.)
***Hint to entrepreneurs: maybe there is a killer opportunity for a risk-free service that aggregates long tail providers (or serves many of them in a harmonious way as a white-label SaaS service). Being more hungry, they will be more willing to innovate on pricing models than the dominant fat cats.
---
*Your feedback is invited. Comment here or email me (FairPay [at] teleshuttle [dot] com).
This post was first published 3/26/19 on my blog at FairPayZone.com.
Minor revisions and enhancements are included in this version (latest: 6/24/19).
------------------------
More about FairPay
(FairPay is an open architecture, in the public domain. My work on FairPay is pro-bono. I offer free consultation to those interested in applying FairPay, and welcome questions.)
A brief introduction is in Techonomy, "Information Wants to be Free; Consumers May Want to Pay"
