TechBio KPIs: Is Your Discovery Platform Validated?
Pablo Lubroth is an investor at Hummingbird Ventures, a global early stage fund, where he covers bio + health. Viswa Colluru is the Founder and CEO of Enveda Biosciences, a techbio company engineering new drugs from nature.
Performance metrics are commonplace among tech companies. They allow internal and external stakeholders to understand how well the company is doing in an objective manner. In early-stage TechBio companies, due to the differences in technologies applied and evidence required to create a successful product, metrics are anything but like-for-like.
In SaaS companies, Key Performance Metrics (KPIs) used to track performance include Annual Recurring Revenue (ARR) growth, retention rate and Customer Lifetime Value to Customer Acquisition (LTV:CAC) ratio. These indicators are useful for the CEO and managers to understand what is working and where they should focus their attention. Likewise, for investors performing due diligence, these metrics give a quick insight on how the company is performing against its peers. The caveat is that these are simple heuristics and don’t guarantee the success or failure of a company. Instead, they can be seen as a guiding principle of what a company’s milestones might look like at different stages of its lifecycle.
In contrast, drug discovery TechBio company performance only becomes more quantitative once it has late-stage preclinical or clinical assets in its pipeline. Therapeutic platform companies that start from zero, however, don’t have a standardized measure of performance. This is not a surprise, given that these platforms can be extremely different from one another. This can create confusion amongst founders on the milestones that the company should be aiming for to “validate” its platform. This grey area leaves founders scrambling to speak to as many investors as possible to understand what “investors are looking for” at a certain fundraising stage. Of course, given the lack of industry standardization, opinions vary from investor to investor, doing little to clarify what success looks like for the founder.
The true north star for any drug discovery company is to provide medicines to the patients. A distinguishing feature of TechBio companies, however, is to not only get drugs approved, but to do so with increased efficiency. The shift in attention to scalability of drug discovery has brought a focal change to metrics besides pipeline progression.
This is clearly exemplified by the S-1s filed by companies ahead of their IPO or in their investor reports. For public biotech companies such as Juno Therapeutics, MyoKardia and Sana Biotech the focus on validation was around proving novel modalities and targets. Public TechBio companies not only have to prove progress in its clinical pipeline, but also in alternative metrics such as library sizes and decreased cost to IND. Below are some examples of Recursion’s and Exscientia’smetrics they exhibited within their reports in addition to pipeline progress:
Intro to TechBio KPIs
Given the number of metrics a company could report on, the key question that arises for founders is “how do I show that my platform has been validated?” Let’s take a look at the types of validation metrics that a company could be aiming for:
Specific: Unique characteristics of your technology, team, or application. These can be qualitative or quantitative but can only be rigorously compared internally.
General: Applied metrics that can be compared head-to-head to others in the industry, independent of the specific technology or platform.
Technical: Indicative or generally accepted proof-points towards creation of commercial value even though they do not fully derisk eventual favorable outcomes.
Commercial: Recognizing these is simple — someone has paid you money and/or contractually agreed to pay you in the future.
If we represent this visually, we get a Validation Matrix that looks like this:
The adage “all models are wrong, but some are useful” is also relevant for metrics. In particular for the Technical column, which are all proxies for i) proximity to drug approval and ii) drug discovery process efficiency. Therefore, the utility of each metric depends on how well it correlates with downstream value that can only be read out many years later (e.g. drug approval or total revenue). For example, quantifying the amount of data generated might be useful, but it is a far removed metric from increasing likelihood of drug approval or improving efficiency. Whereas ‘decreased time and/or money from hit to lead’ can be a useful marker for improved efficiency
Deciding on markers of validation
Your choice of metrics will determine how resources are allocated and how investors will measure your performance. Selecting the right ones can save you time in the long run. Here are a few suggestions on how to find adequate markers:
Use a totality of evidence: There isn’t one perfect formula; you must weave the story using a totality of evidence (i.e using all the available data). This is more critical early on, when you don’t have a “big win” in any one column or are just getting started with an idea. As your company matures, you want to ensure that you are making progress towards the upper right corner of the metric square — ultimately moving towards positive cash flow and profits.
Educate future investors: Starting from the very first pitch to having an investor on your Board, talk about why certain metrics make sense for you. Then use the opportunity to showcase your impressive rate of change. For example, the rate of proprietary clones generated, or the size of your library of unique chemicals, or the number of imaged cells. It should be apparent from your story on how these bridge to short or medium-term commercial milestones on the right-side of the square.
Align with current investors: Once you have a set of metrics that make sense, dedicate a meeting to brainstorming with your Board and align with them. You can then use the outcome of this brainstorm to quantitatively (and without bias) communicate your technical progress.
Let your desired business model dictate your priorities: As you decide the right business model for your company based on your own vision, your understanding with early investors, and the evolving market signals, don’t be afraid to take the shortest path to it. For example, Adimab and Atomwisechose prolific partnership strategies in the biologics and small-molecule domains respectively. On the other hand, Kallyope chose to develop compounds internally and quietly launched 4 clinical programs. Recursionchose the hybrid approach and is pursuing different business models for different therapeutic areas. At this point in time, the public markets seem to largely value pipeline maturity and ownership over technology collaborations, but that may change in the future. Everyone is still learning when it comes to realizing and recognizing value in this space. For now, choose aligned investors and convince them of your path, and the atypical choices you will make on your road to building a highly valuable company.
Pick your tradeoff wisely: “Startups die of indigestion, not starvation” Meaning: you have a very small chance at checking off all the previous suggestions. If you try to, you will likely end up failing at achieving even one of these. Here are some obvious, and some not-so-obvious, trade offs:
Grants might delay operational and technical progress: Grants can be good early validation, and at the right time, provide significant non-dilutive funding. However, beyond writing up your proposal and responding to comments, which can take up weeks in itself, administration of grant awards can be a large burden on a small team. Not to mention that notification of awards can be slow, and the uncertainty can be difficult to plan around.**
Partnering on your platform may delay internal asset discovery: If you think of your platform as a utility, then any “uptime” used by partners or external teams comes at the expense of use by internal teams. It also prevents critical technology development that you most likely must accomplish. Consider partnering on pre-existing datasets, or early assets, or when your technology bandwidth is not limiting to your internal asset development.
Pursuing platform scaling might inhibit pipeline development (and vice versa): In some ways, this is the classic tradeoff that gets the most attention in techbio. When is the right time to “scale” your technology or dataset? Do it too soon and you sacrifice the ability to take your assets further along in preclinical or even clinical development. Leave it too late, and it might tempt your board and other stakeholders to turn your company into a one- or few-asset outfit. Consider a balanced approach, and most importantly one that is aligned with your current and near-future target investors. Over time, you must achieve platform Defensibility Through Execution while building a robust pipeline. The only question is in which order, and with what step-size for each.
Case Study: Enveda’s Dashboard
To bring all the above concepts into practice, Viswa Colluru, Founder and CEO of Enveda, shares his approach to platform validation:
What does Enveda do?
At Enveda, we are uniquely enabling drug discovery from natural products. Using metabolomics and machine learning, our platform indexes novel chemical space at an unprecedented rate. Combining that dataset with dozens of innovations in robotics, screening, fractionation, and in vivo experimentation, we are annotating high-potential leads at scale. On the asset development side, we have built a vertically integrated preclinical operation from Day 0, allowing it to prosecute internal opportunities with high-efficiency; even as a small company. Given the extensive moving parts, we have built a simple, yet effective, set of metrics that we share both internally (with the team) and externally (with investors).
How do you use your Dashboard?
It is a snapshot of the entire technical state of the business. Consider it a tailored, “balance-sheet” equivalent.
The team tracks each of the 7 major steps between discovering new molecules in Enveda’s natural product library and the declaration of a clinical candidate (Step 8). This encompasses both progress through the platform and pipeline. The rate or efficiency of conversion of each is a fantastic proxy for how well aspects of Enveda’s technology, such as algorithms or robotic throughput, are improving over time. The proximal goal is to decrease overall time and effort from Step 1 to Step 8. The subsequent goal is to track and improve the translatability of these compounds to demonstrate differentiated clinical success rates; what matters most.
When the data is viewed over time, the results show the rate of progress for Enveda. For example, as our screening platform successfully scaled over the last 6 months, we saw an exponential increase in our ability to identify novel sources of interesting biological activity.
When paying specific attention to the transition of progress from the platform metrics to the pipeline metrics, you can learn about the rates of conversion of platform substrate to commercially accepted pipeline proof-points.
How did you choose and use validation metrics?
Right after the Series A funding round, the Enveda leadership team did two important things:
identifying pipeline goals for the next 2 years
explicitly mapping key steps in the Enveda process.
This led to the first version of the set of relevant metrics, which was then refined with the help of everyone in the company. Once the final version of the proposed metrics was in place, it was included in the agenda for the next board meeting for alignment in terms of rationale and future reporting.
Today, the stepwise metric model is an integral part of Enveda’s workings. It is used to measure progress toward validation and value-creation, to prioritize different initiatives, and to plan resource allocation and goals:
What lessons have you learned in the process?
Have one set of metrics for both internal and external use: Use the exact same data to rally internal teams, assure your board, and excite your future investors. This way, everyone is aligned on what progress looks like, what progress has been made, and how you create value.
Have a qualitative narrative for each metric: This can be an elaboration of key movement in some metric, or entirely new data behind one of these numbers (e.g., new exciting in vivo data on one of your programs). Not all targets, leads, or screens are created equal.
You get what you measure: No set of metrics, goals or milestones, however well-thought out, are perfect. Be aware that your team does not unintentionally over-index on the numbers or tries to nudge a metric at the expense of others. Read John Doerr’s excellent work on this topic for additional inspiration.
In conclusion
There are no standard metrics for TechBio. We are all learning as we go, but there are some great examples to learn from already. Use this flexibility to your advantage.
We present a “Validation Matrix” that allows you to record and compare both technical and commercial progress in a manner that is specific to you or generalizable to the rest of the industry. Use this to weave the narrative of your validation.
You cannot, and should not, try to pursue all the various types of validation. Let your business model drive your choices. Make sure your stakeholders are aligned with your choices.
Develop a set of metrics that bridge your specific approach or technology to generalizable markers of value. Again, make sure your stakeholders are aligned with your choices. Use these to demonstrate progress in an unbiased manner.
No set of metrics are perfect on their own. Use your judgment while choosing tradeoffs, caution to avoid mis-prioritization internally, and a compelling qualitative narrative to complete your story.
If you’re an early stage founder of a bio company, please get in touch — we’d love to chat: pablo@hummingbird.vc | viswa.colluru@envedabio.com
Thanks to Adam Goulburn, Xinchen Wang, Nikita Andersson and Gunnar De Winter for all your great suggestions.
**A note on grants: consider nominating a dollar-value to your team’s effort after a value inflection event, such as a fundraise or key data. If the Adjusted Grant Value is not positive, defined as the expected value of the grant minus your team’s effort, move on to the highest leverage efforts. [Adjusted Grant Value = EV — Team Effort, where EVgrant = Pgrant * $grant]
 | A guest post by
|