Episode 20 of Silicon Valley explores numerous classic conflicts: consumer-oriented v. enterprise business model; engineers v. sales; revolutionary vs. safe; long-term goals v. short-term profits; Erlich v. Jian-Yang … Richard’s plan has been to transform the world by giving away the basic version of Pied Piper’s revolutionary compression technology, rapidly building a huge user base, and hoping to charge for premium features one day (the consumer, “freemium” business model). However, CEO “Action” Jack Barker and his new sales team—preoccupied with implementing his “Conjoined Triangles of Success”—want an immediate focus on revenue and insist that Pied Piper make enterprise software they can immediately sell to big business customers. Worried even that won’t be easy enough to sell, Jack’s team strips away every cool and revolutionary feature until they’ve transformed Pied Piper into a business-facing appliance company selling Pied Piper Boxes that Jared deems “rectangular, glorified thumb drives that resemble nothing so much as old Betamax machines. ” Richard is horrified and dismayed.

But is Jack’s sales-driven, enterprise business model really that bad? After all, this plan is to sell Pied Piper’s Boxes—to a few, large businesses that can pay more than individual consumers ever would—and generate an actual revenue stream now. But targeting the enterprise market has its difficulties. For one thing, big business customers are likely to require a lot more security, reliability and support. Early customers can wield tremendous power to demand features and custom solutions. Focusing on reacting to these demands could lead to incoherent product development, and cause Pied Piper to abandon long-term strategies, including any ‘risky’ cutting-edge projects. In other words, by putting all its energy into meeting current customer expectations, Pied Piper could sacrifice its ability to meet future demand, since (as Steve Jobs famously pointed out), customers don’t necessarily know what they want – or what they will want – sometimes you have to show them. Sacrificing the cutting edge could be especially dangerous since the ex-Nucleus team looks poised to jump into the market and capitalize on any misstep by Pied Piper.

But let’s assume Pied Piper sticks with Action Jack’s enterprise business model. To get under way, Pied Piper will need, at minimum, a well thought-out license agreement, service agreement, or other governing customer contract. There are many things this document should contain. For one thing, Pied Piper doesn’t want its customers to license one Pied Piper Box and then clone the software to avoid paying for more Boxes, and Pied Piper really doesn’t want someone to reverse-engineer or decompile the code and start competing with Pied Piper. (After all, that’s how Hooli stole Pied Piper’s first-gen algorithm and established Nucleus!) I’ve previously discussed how a license agreement might have helped Pied Piper keep control of its technology and fend off Hooli’s lawsuit. Pied Piper wants its new contract to get this right.

Second, Pied Piper’s target business customers will try to impose exacting security, reliability and support requirements. Pied Piper wants its contract terms to limit its obligations in these areas to what it can actually deliver—provided the sales team hasn’t already blown it by promising customers the moon—which looks like it could be a recurring problem under Action Jack’s sell-it-first, build-it-later strategy. Related to that, Pied Piper also wants a “limitation of liability” clause that caps how much Pied Piper has to pay an angry customer if something goes wrong (like a Box catches fire, or a bug prevents de-compression, trapping the only copy of some amazingly valuable content in an unusable format). Pied Piper might not be in strong enough bargaining position to demand this, but suppliers often seek to limit their liability to the price of the product, or to the amount the customer paid under the contract in the prior 12 months, or something similar. In any case, Pied Piper will want to eliminate certain categories of damages all together, and limit its total liability as much as possible.

Pied Piper also needs to think about how long to support any particular version of its Box, and how to ‘force’ customers to upgrade to new ones on Pied Piper’s schedule. (It can be a headache when customers won’t let go of the ‘old’/unsupported version, and Pied Piper may want its contract terms to help nudge customers along.) Pied Piper also may need to obtain contractual authorization to use customer data in certain ways, for example if Richard salvages the plan to use a neural network and artificial intelligence to learn from customer data and improve the algorithm. Finally, many large corporate customers will require negotiated and customized contract terms. Pied Piper may have to hire an in-house attorney to handle this, or spend a lot of time on the phone with Pied Piper’s outside counsel Ron LaFlamme.

But maybe Pied Piper won’t actually abandon Richard’s consumer-oriented, freemium business model. Sure, there’s no hope his model will generate revenues anytime soon. In fact, a big risk is that Pied Piper might never figure out how to monetize it. But it has the potential to gain millions, even billions of users and revolutionize the internet! So what would Pied Piper need if it went down this path?

First, this business model puts Pied Piper in much different legal environment. Pied Piper will be handling—on its platform, servers, cloud—all sorts of user generated and uploaded content. As I described in this post last season, there are some basic ways in which Pied Piper can protect itself from legal problems caused by its users’ content. It is especially important for Pied Piper to think about these issues before it allows users to share content through its platform.

Second, Pied Piper absolutely needs a privacy policy and terms of service—and to make sure these documents are binding on users, it should require users to “affirmatively” agree to them (with a check-box or click-through) when they sign up for Pied Piper’s service. Why these documents? For one thing, California law requires a privacy policy for any website or app that collects user ‘personal information’ (which is broadly defined to cover names, email addresses, and probably even IP addresses or cookie data). The privacy policy should explain what kinds of personal information Pied Piper collects, how that information is used and shared, and a few other things.

Pied Piper will want to include a variety of provisions in its terms of service. I don’t have space to discuss them all, but a few things leap to mind. For example, as with corporate customers, Pied Piper wants to obtain consent and a license to use customer data in connection with its services, including its neural net and artificial intelligence. This should deter customers from claiming copyright infringement, or claiming that Pied Piper is unfairly profiting from their content. (In case that sounds paranoid, Google was sued for requiring customers to decipher a two-word “CAPTCHA” in order to open free Gmail accounts. Decoding the second word helped Google digitize books and addresses, and the lawsuit essentially claimed that Google was “stealing” its users’ time and attention for Google’s commercial benefit. The court thought the lawsuit ‘defied common sense’ and tossed it out of court. A lawsuit alleging Yelp improperly profited off user reviews met the same fate. It’s never fun to get sued, and Pied Piper would rather avoid any more lawsuits than have to win them.) Pied Piper will also want terms that define acceptable use and let Pied Piper terminate anyone who misbehaves, permit Pied Piper to change or stop offering its services, and limit its liability in the event something goes wrong or it is hacked.

We don’t yet know which business model Pied Piper will choose, but it would be dangerous for it to flip-flop or remain undecided for too long. Not only could Pied Piper burn through its funding and have nothing to show for it, but it would open the door to competitors. And that ex-Nucleus team looks certain to start causing problems, maybe as soon as the next episode.