What a buyer pays for is not how much revenue you have. It is how much of it comes back next year without being re-won.
J Chad Brown, CEO, CommEx Advisors, The Value Multiplier anchor essay, October 2026.
A business can grow its revenue and lower its value at the same time. That sounds like a contradiction, however in tools and diagnostics it is a real risk, and it is one worth monitoring as a metric in its own right rather than discovering later.
It happens because the market does not price revenue. It prices the odds that the revenue repeats. And in this sector the single biggest signal of those odds is the share of revenue that recurs, which is another way of saying the share that comes from consumables rather than from placing instruments.
That share decides which group of companies a buyer prices you against, and the two groups are a long way apart.
What I watched happen
I joined NanoString as SVP of Sales and Marketing in July 2017, into a business that was decelerating fast. Product and service revenue had grown 42% in 2014, then 27%, then 22%, and then 4% in 2017.
The 4% is the number people remember, but the 4% was not the problem. The problem was the shape of the line before it. A single bad year can be explained by timing or by one large order slipping. 3 consecutive years of slowing growth is not timing, it is the way the business is built.
The stall, and what followed
NanoString product and service revenue growth, year over year, against the underlying tools market. The commercial rebuild began in July 2017. Large cap tools organic growth ran at 4 to 8 percent a year over the period, with 2020 and 2021 running well above the band on COVID response revenue.
- 2014: 42%
- 2015: 27%
- 2016: 22%
- 2017: 4%
- 2018: 16%
- 2019: 24%
- 2020: 7%
- 2021: 29%
Source: NanoString SEC filings and earnings releases. Product and service revenue only.
Underneath the deceleration sat 3 things, and they were connected.
- Growth depended on placing new instruments, and placements were falling. In the third quarter of 2017, the low point, they were down roughly 40% year over year.
- Reported growth was being flattered by collaboration revenue, which was 37% of total revenue that year, non-recurring and outside the commercial team's control.
- Every instrument already placed represented years of future consumable revenue, but the commercial organization was built to win new placements rather than to grow consumption in the accounts it already had.
*The work ahead was a strategic shift, from building the installed base to growing the recurring revenue that sits on top of it.*
What we changed
The first move was structural, and it built on a change that was already under way. Earlier in 2017 the company had added consumable reps and an inside sales team alongside the instrument reps. However, the consumable reps and inside sales covered the same accounts, and both were paid on all of the revenue in them, so the company was paying twice for the same dollars and neither role clearly owned the account.
In 2018 we added roles and gave every account one owner. Instrument reps carried approximately 75% instrument and 25% consumables, and consumable reps carried approximately the reverse. Inside sales took the long tail, meaning the many accounts with low individual revenue, while consumable reps concentrated on the highest volume accounts.
That account split was deliberate and it was about ownership rather than efficiency. I did not want two people paid on the same account, because when that happens neither one is accountable for it. Every account had one owner.
Then we went further. Consumable reps moved to 100% consumables, with no instrument component at all.
We also created a Global Consumables Director, who managed the inside sales group directly and the consumable reps on a dotted line. One person whose only job was consumable revenue. And we built the coverage around a small team per territory: instrument rep, consumable rep, inside sales, field application scientist and field service engineer, meeting face to face on a regular cadence so the customer experienced one company rather than 5 functions.
The resistance
Not everyone agreed with the change, and the concerns people raised were fair ones.
The main concern was that a consumable rep with no instrument component would have no reason to help place new instruments, and might quietly work against them. There was also a fair worry about income during the transition year, when people would be earning against a target they had not sold into before.
My answer to the first one was that the incentive is already there, it just runs one step further out. If you are a consumable rep, the best thing that can happen to you is a larger installed base to sell into. You do not need to be paid on a placement to want it.
The concern I did not anticipate turned out to be the one that resolved itself in the opposite direction. We expected some confusion about who a customer should call. In practice the split made it clearer, and it is a common enough structure in mature businesses that customers recognized it.
How long it took
Improvement showed in 2018, the first full year, which is faster than this kind of change usually moves. Growth went from 4% to 16%, then 24% in 2019.
Market conditions helped as well. nCounter sat between qPCR at a handful of genes and sequencing above 10,000, and that mid plex space was genuinely narrow. At the time, though, sequencing platforms did not work well with FFPE samples and were priced far higher, which left real room for a mid plex platform in translational and clinical research. Part of the pull through improvement came from that fit.
That window narrowed later. Sequencing prices kept falling, and the technology became more FFPE compatible, which removed both of the advantages that had made the mid plex position work. Some of the later pull through slowdown traces back to those changes in the competitive landscape, especially as the sequencing installed base kept growing. Illumina alone reported an active installed base of more than 11,000 sequencing systems at the start of 2018.
The pull through number itself moved more slowly, from about $78,000 per system in the net installed base to $80,000, then $87,000. That is the honest pace. Behavior changes in a quarter when compensation changes, but the revenue effect compounds over years, because it depends on accounts working their way up a utilization curve rather than on a single decision.
The case study also shows a higher figure for 2021, about $112,000, but that figure includes GeoMx consumables, so it isn't comparable with the nCounter numbers here. For nCounter alone, the 2021 figure was about $82,000.
Across the 4 years, product and service revenue went from $72 million to $144 million, roughly 19% a year. Much of the growth after 2019 came from GeoMx, our spatial biology platform, which we launched into the same installed base without moving the nCounter team onto it. The underlying market grew at mid single digits over the same period.
The full numbers, and how each one was measured, are in the NanoString commercial case study on our website.
What we got wrong
2 things, and both are the same mistake seen from different angles.
We never scored accounts before approving a placement. There was no test of how much an account was likely to consume before an instrument went out the door and there should have been. We had a strong focus on immuno-oncology accounts over other applications, and that focus was right, but pull through still varied widely from one immuno-oncology account to the next. The application alone did not predict consumption. The highest pull through accounts had other things in common, and that profile would have made a usable ideal customer profile for placements. We never turned it into a guide the field could apply. Compensation changed where the field spent its time, but it did not stop instruments going into accounts that were never going to run them.
And we did not manage dormant accounts well enough. Some customers stopped running their systems, and nobody owned bringing them back. The installed base figures show a related problem. The gross installed base grew from about 600 systems to about 1,050, while the net installed base, on a 5 year useful life, levelled off near 600 and then dipped as the earliest systems aged out. The headline number kept rising while the number that drives consumable revenue did not.
The part we owned was the part we did not measure. If I were doing it again, the first thing I would build is the account scoring, and the second is a far more deliberate regional marketing effort to keep finding new applications for the base that already exists.
Why any of this decides the multiple
Here is the part that turns an operating story into a valuation one.
The largest companies in this sector lead with their recurring share when they report, because that is what the market rewards. In their 2025 annual reports, Thermo Fisher took about 84% of its revenue from consumables and services, Danaher reported about 82% of its sales as recurring, and Illumina took about 74% of its revenue from consumables alone. Those are the businesses priced at the top of the range.
Further down the market the same logic applies. A buyer looking at a consumables heavy business and an instrument heavy one, in the same sector and the same cycle, will pay more for the first, because more of next year's revenue is already in place before anyone sells anything.
You can also watch the market price it directly. Pull through per instrument is a disclosed metric for the companies built this way, and it comes up on earnings calls as a matter of routine.
Illumina has published target pull through ranges by platform for years. PacBio guided to $225,000 to $250,000 per Revio system a year, landed near $202,000 last quarter, and cut the range to $200,000 to $225,000. 10x Genomics has reported pull through per instrument since its IPO filings.
The clearest example is the newest one. Alamar Biosciences went public in April and used its first earnings call to name 2 priorities for the year: adding at least 100 instruments to the installed base, and holding per instrument pull through above $400,000, which management called the clearest indicator of platform utilization at scale. In that quarter, consumables were 53% of total revenue, up from under 40% a year earlier.
Alamar's CFO also explained that new placements will put some pressure on pull through in the near term, because the number of instruments added each quarter is large relative to the installed base. That's a denominator problem, and it applies to any company built this way: new systems that haven't ramped yet pull the average down, and old systems that are no longer running do the same. Alamar chose to explain it to investors on its first call.
Analysts ask about it unprompted, and they ask early. On PacBio's most recent call an analyst asked what pull through to expect in the second half as clinical and population scale customers ramp. On 10x's call an analyst asked whether the previous platform's pull through was a reasonable starting assumption for its new spatial platform. Those are questions about the commercial engine, asked by the people setting the price.
So two companies can ship the same instrument, book the same revenue, and be valued as different kinds of business. What separates them is attach rate, pull through per placement, whether the instrument is actually being used after installation, and whether the commercial organization is managed and paid on placements or on consumption.
Every one of those is a commercial decision. None of them is a product decision. And all of them are made years before anyone opens a data room.
What this means if you own one of these companies
5 things to ask, and none of them require a project.
- 1. Do you know what the highest pull through accounts have in common, so that the profile can be used as a guide for the next placement?
- 2. What is revenue per installed instrument, by cohort year, and is the newest cohort behaving better than the one before it?
- 3. What share of the installed base has gone dormant, and does anyone own bringing it back?
- 4. Is anyone paid on consumption, and is it enough to help them prioritize where they spend their time? If so, what are the top performers doing differently, and how do you replicate it across the other consumable reps?
- 5. Before the next instrument ships, is there any test of whether that account will actually run it?
The first one is the cheapest to fix and the one most companies skip, which is why it is the one I would start with.
How we help
Each of the questions above has an answer that can be built, measured and shown to a buyer. That is the work we do, and it is the same work described in this essay, put in place as one system rather than as a set of separate fixes.
It starts with the Commercial Readiness Assessment. The assessment scores a commercial organization against the 9 pillars of our Commercial Operating System, using 42 statements that each need evidence behind them. It produces a score out of 100, a map of where the gaps are, and a 90 day plan ranked by what moves the score most. Run at diligence, in the first 100 days, or as a portfolio baseline, it gives the owner and the management team the same view of the commercial engine.
On revenue mix, the work covers 7 areas:
- Account scoring before placement. We build the profile of the accounts that actually consume, from the company's own pull through data, and turn it into a test the field applies before an instrument ships. As the immuno-oncology experience shows, the application area alone is rarely enough.
- A named owner for utilization. One person accountable for consumable revenue across the installed base, with dormant accounts tracked and a plan to bring them back.
- Coverage and compensation pointed at consumption. One owner per account, and a pay plan that rewards consumption, which is what a buyer is paying for.
- A cohort view of pull through. Revenue per installed instrument by placement year, measured on the net installed base rather than the gross count, in a form that can be shown to a buyer.
- Forecast discipline. Deal qualification on MEDDPICC and a weekly forecast cadence, so the number the team commits to is one the board can rely on.
- Business reviews built on shared dashboards. Quarterly reviews that put Sales, Marketing, Customer Experience, Finance and the senior team in front of the same numbers, rather than each function reporting its own version.
- Sales process and training. One sales process, used by everyone and inspected by managers. Which process you pick matters less than whether it is used consistently.
Each area is put in place the same way: a named owner, a defined way of working, the tools and dashboards to run it, a regular operating cadence, 4 KPIs that show whether it is working, and a path from where the company is today to where it needs to be. We then re-run the assessment against the same 42 statements, so the improvement is measured rather than asserted.
None of this promises a multiple, because the market sets the multiple. What it does is remove the specific discounts a buyer would otherwise apply: revenue that depends on new placements, an installed base nobody is working, and a forecast nobody can defend. Each of those is documented in diligence, and each can be fixed well before anyone opens a data room.
Measure it before it hurts. Book a conversation at www.commexadvisors.com
Sources
- NanoString Technologies: Forms 10-K and 10-Q, 2014 to 2021, including the Q3 2017 Form 10-Q; NanoString commercial case study (CommEx Advisors).
- Thermo Fisher Scientific: 2025 Form 10-K (revenue by consumables, instruments and services).
- Danaher: 2025 Annual Report (recurring and nonrecurring revenue by segment).
- Illumina: 2025 Form 10-K; J.P. Morgan Healthcare Conference remarks, January 2018; Q4 2019 earnings call.
- PacBio: Q2 2026 earnings release and call.
- 10x Genomics: IPO registration statement (2019); Q2 2026 earnings call.
- Alamar Biosciences: Form S-1 (2026); Q2 2026 earnings call.