2nd Sep 2020

Guide through license-royalty models for early-stage founders

Over the past couple of years, I’ve had the privilege of sharing Boards with some of the most experienced former executives in the licensing world such as ARM and Imagination Technologies. License-royalty business models can be a lucrative way for tech entrepreneurs to build large and valuable companies. Other than being great companies in their own right, nuTonomy, Quantenna, Swiftkey and many others also grew by licensing their technologies.

In this post, I will share some of my learnings and dive into the common negotiation pitfalls early-stage companies should watch out for; ranging from procurement and pricing models, through to working capital considerations. Not got time to read this all? Here are the 6 key takeaways:

  1. Long-term defensibility and building value perception to the buyer is critical, particularly in consumer electronic products
  2. Beware of the tech giants when licensing to them, their procurement and legal departments are ruthless
  3. Whether it is a tech-giant or an electronics manufacturer, make sure you have the “ability to audit” the end-devices sold and try to get an umbrella deal
  4. Seek local experience when negotiating, expect friction on upfront terms and beware of advance royalty pitfalls
  5. Be prudent in your forecasting and cash management and beware of potential annual caps at scale
  6. Define volume calculations carefully — ideally, elect to use ‘shipped’ vs ‘sold’ or ‘online’ end-products

Before we start — what is a license-royalty model?

As an example, a car manufacturer designing a new electric vehicle would choose to build some components in-house (e.g. body, gearbox, interior) but they will need to buy-in many other components (e.g. recharger, battery, parts of the operating system etc.). The need to buy-in is often to incorporate the most state-of-the-art technology, enabling them to compete with others in the market. The same goes for hardware and software subcomponents of consumer electronics, mobile manufacturing and many other sectors, right down to the semiconductors themselves. License-based technology companies are beneficiaries of this.

For starters, license-royalty models can come in lots of different shapes and sizes, but typically, they comprise four key components:

Non-recurring engineering (NRE) costs — paid upfront (on signing) and assigned to integrations, product enhancements or other ‘one-off’ engineering work.

License Fee — the right to access the technology for development and/or production. Typically on a per-use basis or multi-use for a period of time.

Advance royalty — paid upfront (typically after you have integrated/delivered the technology) by the licensor and offset against future royalty payments. Can be known as the ‘pre-buy’.

Royalty schedule — based on the royalty agreement, these are periodic payments made on the usage of the technology over time, typically a price per unit sold.

Support/Service fee — cost to help support the customer and is usually a % (10%-20%) of the initial license fee. It also recurs annually, but some customers choose not to renew after the first year.

A license-royalty model can deliver massive scale — but it’s harder than ever. Here are 6 tips to help you on the way:

  1. Long-term defensibility and building value perception to the buyer is critical, particularly in consumer electronic products

    10–15 years ago, it was much easier to put together a license agreement as it was much easier to show differentiation. Your proposition has to have long-term defensibility and be non-replicable. Due to the rapid advancements in the software world, it has become more difficult to build and protect a piece of IP. You typically have a 2/3 year advantage before others catch up.

    Commoditisation typically drives margins down for consumer electronic products — this puts pressure on license agreements (usually 3–5 years long). Therefore, you need to understand the entire end-product cost, not just your segment or sub-segment. For example, in embedded voice recognition licensing, if the license is a proportion of the whole end device royalty stack, procurement will look to bring this down below a certain threshold as an entirety. As such, you are left competing with other parts of the bill of materials and are left trying to build value perception to the buyer relative to entirely different technologies — so be prepared for this. The proportion you get when you start is likely to be the best-case and will erode over time. As a data-point, embedded predictive text licensing margins halved over ten years.

  2. Beware of the tech giants when licensing to them, their procurement and legal departments are ruthless

    With many of the giant established tech companies (e.g. Google, Apple, Microsoft), they do not have a licensing framework as licensing is not in their DNA. They instead have early acquisition strategies, an option that is often of no interest to early tech entrepreneurs who want to build large standalone businesses. For many teams you will deal with within these organisations, it will be the first time they will have ever licensed a piece of technology. In these cases, one of two things happen: either the sales cycle lengthens significantly, or the technology is licensed without a proper licensing framework in place (usually this means the company is desperate for the tech!).

    Procurement & legal departments at the tech giants are robust. Many of the terms you ask for are stricken out, and you have to choose your fights wisely. It is essential to include wind-down terms: if a company discontinues a product and thus terminates the contract, there should be terms for the time it takes to wind-down sales (as this can continue to be significant). In fact,

    "there have been several instances whereby tech giants have killed/pulled out of deals post signing a contract, post paying and even post-launch if market feedback is poor"

  3. Whether it is a tech-giant or an electronics manufacturer, make sure you have the “ability to audit” the end-devices sold and try to get an umbrella deal

    Original equipment manufacturers (OEMs — companies such as Airbus, Dell, Samsung, Sony, Toyota etc. that provide components in another company’s product) usually think of products in generations, whether this is a chip, a phone or an electronic device. When negotiating with these large OEMs, umbrella deals (which include a “group” of current and future products) are more favourable than individual device deals. This avoids yearly re-negotiations with procurement departments whereby they re-visit many of the terms previously agreed.

    An OEM will never want to give you exact or projected volumes; in many sectors, there is a lack of transparency in the market on volumes due to competition. Though auditability is common practice (de-facto standard for most OEMs), this term is usually insurance for the licensee and not often enforced. We do still see many of the large OEMs push-back on the ability for a company to audit them contractually; you should try to hold firm on this point.

    "Without over-generalising, there is evidence to suggest that many large companies do understate volumes of end-device sales and many industry veterans would say that this is a ‘cost of doing business."

  4. Seek local experience when negotiating, expect friction on upfront terms and beware of advance royalty pitfalls

    NRE is paid upfront and can make up a large proportion of the total contract value; up to 50%+ in some cases. Entrepreneurs should seek to be creative and anticipate the time needed to integrate (and some times further develop) their technology.Lawyers for licensee always want to own some of the resulting IP of their NRE, and start-ups should be careful of and avoid this, as it usually puts them on the back foot when negotiating. As the technology matures and integration requirements reduce, you may want to incorporate NRE into the license fee to avoid this point of friction.

    The license is then negotiated per-use or multi-use for a specified period. Although this can vary whether a company is in the ‘design’ or ‘manufacturing’ phase of their product, typically, this is a subscription fee that is paid annually for 3–5 years. Vendors often provide a multi-year upfront option with a discount to help improve cashflows. ARM as an example, had 60–70% of revenues recurring each year owed to licensing as they grew, which is what I like to think of as ‘secured’ top-line.

    The advance royalty (or “pre-buy”) is dependent on what the licensor is comfortable with and is often a key point of negotiation. Founders must understand that an advance royalty is an advance on the royalty schedule and should not impact other parts of the deal. Great negotiators who know their volumes will be high would be happy to inflate the advance royalty payment at the expense of getting a discount on NRE or the license costs — this is usually a sign that they are confident about their volumes. It may look great because you’re getting lots of money upfront, but often the total deal value is much lower:


    Western buyers are more comfortable with an upfront payment, but there is no particular rule-of-thumb in the market today on specific proportions vs royalties. Buyers in the Far East don’t tend to like advance royalties. So local relationships or experience within the management team is key to negotiating and doing sustained business in the region.

    The royalty schedule then kicks in once the pre-buy is used up and is typically banded. The more mature and established markets (e.g. predictive text — brownfield market today) usually have 4–6 tiers based on the number of devices, e.g. Band one: 0 to 1 million, Band two: 1 to 3 million, Band three: 3–5 million etc. The higher band pays lower royalty values, as expected, ramping down to a ‘floor’.



    Many companies tend to overestimate their volumes, and so will focus on the ‘floor’ in the negotiation phase. Though they quite often never get there. The key after this is to license new technologies to the same customer and hence reset the royalty ramp down.

  5. Be prudent in your forecasting and cash management and beware of potential annual caps at scale

    Most companies choose to report quarterly (however, this is negotiated upfront and can vary). Working capital cycles can be long, so ensure you have the balance sheet to endure waiting times. There have been some instances whereby OEMs report 60 days post-quarter-end, then pay 45–60 days post receiving an invoice. Push for 30 days where possible.

    Forecasting is also challenging, predominantly due to the lack of near-term visibility. As such, monthly and quarterly restating is common practice for most licensing companies. There are some sectors such as the semiconductors whereby known product cycles can support forecasting; however, there is an expected margin for error.

    As you scale, one of the critical points of negotiation is the idea of an annual cap. If you become large enough from a volume and price perspective, large OEMs can apply a cap on the number of yearly royalties they pay you. They’ll say:

    "No matter how many products we sell, we’ll pay you $Xm per year in royalties"

    For us investors, this is a crucial point of consideration when thinking about the total addressable market as simply multiplying devices by royalty no-longer becomes an accurate enough representation, capping the upside. By this point, you should have an established team of negotiators to fight this off. Negotiations often lead to acquisition offers if an OEM decides they can’t live without your product but will be paying you too much over time if they continue with the pre-existing terms.

  6. Define volume calculations carefully — ideally, elect to use ‘shipped’ vs ‘sold’ or ‘online’ end-products

    The basis by which you count device volumes is typically ‘shipped devices’. Avoiding conflicts around returns, activation timings, user set-up requirements etc. However, some contracts I’ve seen more recently count volumes on activations, usually meaning an end-user buying and using the product or even electing to utilise the additional functionality in more extreme cases — try to avoid these. Within the shipment terms, there are many other terms included, such as the exclusion of returns & inclusions of prototypes and demos, which are relatively common.

    Shipment numbers tend to lead consumer adoption. However, shipment numbers can be up to 20% higher than activations, mainly due to industry ‘grey market’ effects. As crazy as this may sound, take mobile phones as an example, it is not uncommon for there to be many units peeling off or sold in the black market. Products can also sit on shelves for long periods which also contributes to the difference.

The largest and most successful companies are those who can build intertwined relationships and an eco-system around their technology. Supporting and making it easier for your eco-system to use your product goes a very long way in the licensing world, and you could become the industry-standard, like ARM have done in digital IP.

It’s worth saying there’s no one-size-fits-all approach when it comes to license-royalty models. Some of these tips will work for you, and your customer and market will dictate others. Nonetheless, seek experience and support early in your journey to give yourself the best shot at success.

Question or comment on the above? Got a start-up you want to discuss? Do reach out!

If you’d like to read our next post, please subscribe to our newsletter.

Special thanks to Sir Hossein Yassaie, Pete Hutton and Robert Swann for their invaluable contributions to the post.

Newsletter Sign Up

Sign up for latest news