Third Bridge专家访谈讨论液相色谱仪交付验证与Agilent运营节奏，未形成明确交易催化

[[00:01:33]]
What standard operational milestones need to be achieved before a customer lab can formally sign off on newly delivered LC [liquid chromatography] system fleet?

[[00:01:50]]
Just a question. Are we talking about liquid chromatography, or LC, or LC-MS?

[[00:01:57]]
LC, but if we can delineate across LC-MS afterwards, that would be great, too.

[[00:02:02]]
I was just asking because LC-MS is more critical than LC. Answering to this question, the company or let's say, a site needs to go through the sequence of PQ, IQ and OQ in order to validate the output of the measurement validate that the output of the new platform, there is equivalency with the old platform. Here, we are talking about new hardware vs a routine platform upgrade. New hardware is a new platform. Typically, from a, let's say, qualification point of view, from, how can I say, the provider, it's not a lot. It's 3-5 days. The provider to set up, install an instrument, guaranteeing functionality against the factory specification. The typical time for an end customer to run these evaluations typically is 30-60 days. I'm saying typically because these 30 days can be, let's say, reduced if there is a rush or if the project is not a priority can extend maybe to 90 days, but typically, it's 30-60. That's the timeline.

[[00:03:53]]
Just to confirm, that's for LC and not LC-MS [liquid chromatography-mass spectrometry]. Is that correct?

[[00:03:59]]
I would say, yes, we are going through LC. LC-MS is similar situations, just the timing is extended, let's say, plus 25-30% of the previous numbers simply because the LC-MS is a more complex instrument than LC.

[[00:04:27]]
How drastically do the site qualification timelines vary when deploying an entirely new hardware generation vs a routine platform update?

[[00:04:50]]
Let me say, the 3-5 days installation time is the same. This is fixed, it's not variable. The end customer qualification time is the variable. If we are talking about a new hardware generation vs a routine, the new hardware generation typically is what I said before, 30-60 days time for an LC in order to be evaluated comparatively evaluated against previous hardware generation platform.

[[00:05:36]]
Does an enterprise-wide hardware roll-out typically stall if performance variations or configuration bugs emerge immediately after shipment delivery?

[[00:06:13]]
Typically, the deployment of new hardware generation instruments are evaluated before. There is either a situation where an instrument is provided by the instrument provider, like a demo-type of situation, so that the evaluation of the platform can happen before placing an order. It's unlikely that there is an order that is placed without any pre-evaluation done. The method to make sure that the rollout goes smoothly is to run a pre-evaluation. If there is an issue after shipment, meaning that there is a deviation between what has been pre-evaluated and what has been shipped, it depends. It depends what the issue is. Typically, the needle is moved based on the quality of data. Is the issue affecting the quality of the reading of the instrument? Yes or no. If yes, that's serious. If no, let's say, a lamp that doesn't switch on, that's, let's say, not a big issue.

[[00:07:54]]
What variations or software integration issues are testing laboratories running into during initial next-generation LC configurations?

[[00:08:08]]
The situation is similar. If there is a software evaluation, a new platform with… Yes, let's call it, platform with a new software. The evaluation is done in a comparative way. If there is a situation of a defect found on the instrument, the principle is the same. Is a defect affecting a functionality, a peripheral that doesn't connect. Okay, there is a software to be changed to be updated. That's not a serious issue. Is the issue involving the quality of the reading, reproducibility of the reading of the instruments, that's a severe issue.

[[00:09:06]]
Are the early application and hardware performance variations directly translating into customer installation sign-off delays or any unexpected near-term equipment revenue shortfalls?

[[00:09:27]]
Yes. If there is a situation of deviation of reading against a comparative situation pre-deviation, then the fulfilment machine engine is compromised. Now, if we are in a situation of hardware, it's more difficult to recover. If we are in a situation of software, it's easier simply because software can be upgraded easily than hardware replacements. I would say, in terms of revenue recognition, in the market, the revenue recognition is typically at shipment level or receiving from the end customer. In the life science market, it's unlikely to have a revenue recognition situation that is, let's say, outside the so-called Incoterm. Yes. I'm talking about the private market. If you are talking about the academia government market, it's the installation and qualification and validation, it's more likely to be a milestone that affects revenue recognition, but in the private instrument in this market, the revenue recognition is not an issue. The issue is the customer has a problem, then it needs to be fixed, but it's not affecting revenue recognition.

[[00:11:17]]
When a lab encounters configuration friction points with a newly launched hardware line, under what conditions do you think it will freeze active multi-unit upgrades and reroute capital budgets to direct tool competitors?

[[00:11:34]]
That's, let's say, very unlikely situation to happen. The shift of CAPEX budget from, let's say, vendor to vendor under the same project. It should be very unlikely to happen. I don't recall situations like that. The impact of delaying a project is there, but the shift of capital from vendor A to vendor B in a project like that, very unlikely to happen, even though the situation would be severe from a technical whatever point of view.

[[00:12:29]]
Given that pre-order demos usually insulate the aforementioned accounts, are there any specific baseline failures in a new fleet that it would take to cross the extreme threshold where a customer breaks a legacy vendor relationship?

[[00:12:56]]
Just to make sure, we are talking about replacement with new technology in a large-type of installations? We are not talking about a single lab with one unit?

[[00:13:15]]
Yes.

[[00:13:16]]
So if we are in that situation, there are methods to pre-assess in advance. The typical method is demo, send a demo unit to the customer. It's pretty common. In the instrument life science market, there are plenty of demo units going around, if the demo is not a possibility, then there are two avenues. One avenue is to run some evaluations using the company's called COE, Center of Excellence, where a customer can have access and can use the instrument located in the company's Center of Excellence for running experiments, whatever. Of course, this is not a situation where an instrument is used for 30-60 days, as I was saying before, but for a reduced period of time. That's one method of anticipating a rollout and a purchase and a rollout of new technology. Another method is to buy a sample unit that eventually can be kept in production or eventually can be kept as, let's say, extra unit, golden unit at the customer premises. Running the processes on the golden, call it, unit and based on the result of this run, determine the rollout of the, let's say, upgrade of the other units or expansion or whatever.

[[00:15:12]]
What operational and facility hurdles dictate the lag time between a signed instrument purchase contract and the final top-line revenue recognition?

[[00:15:29]]
That's an interesting question because we are in a market where the time between placing an order to the time of fulfilling an order is pretty short. It can be from 4-6 weeks up to three months. From both the customer and the provider point of view, both entities work in this market type where the timing is what I said before. We are not talking about a long time. I don't see this as a big issue.

[[00:16:25]]
Are there any customer facility readiness parameters or regulatory compliance workflows that add the most amount of time to the post-delivery loop?

[[00:16:39]]
This is a variable. The customer readiness is a variable because in a lab, there is an infrastructure to be set up in order to perform an installation of an instrument. Now, if we are talking about an existing lab, probably the set-up is already almost there. If you are talking about a new set-up, there are a lot of variables in order to have the laboratory or the facility, let me say, ready, so at the right time. Typically, yes, one of the variables in moving times in these projects is the infrastructure readiness. There are many things that need to be ready in case of a new infrastructure set-up. Yes, that's typically a major source of mismatch of the project time and then the actual time. From a regulatory point of view, it's embedded in the 30-60 days stated before. If there is a situation of, let me say, equivalency, the instruments will not measure the same thing, but if there is a situation of equivalency, it can be resolved within that period of 30-60 days. If during the period, 30-60 days, then there is something that does not match with the concept of equivalency, then it can be an issue from a regulatory, sorry, point of view, but very unlikely to happen, based on my experience.

[[00:18:54]]
If validation is delayed on the bench, does that essentially trap customers' working capital or delay follow-on consumable orders?

[[00:19:23]]
Working capital, I would say, not much in the sense that the installation timeline can be replanned typically, can be replanned due to the fact that we are in a situation of delivery times that are not extra long. From a working capital point of view, the impact can be mitigated, let me say. On the second point, the consumable, the longer the project rollout is the later the consumable flow will happen. Now, if we are in a replacement-type of situation, there is no impact because the current set-up is using consumable. The new set-up will use consumables. It's a shift, let me say, of use of type of consumable. If we are in a new set-up, then yes. A delay in the project will imply a delay of using consumable.

[[00:20:38]]
When Agilent maintains an instrument book-to-bill ratio at or above 1x over two consecutive years, does that baseline track an expanding backlog of unfulfilled equipment orders or just an accelerating pace of physical factory conversion?

[[00:21:00]]
Let's distinguish if the numerical value is 1.01 or if it's 1.1, it's a totally different story. Let me say, if we are in the last 10% range, the set-up in instrument manufacturing and I want to say it's a set-up in the industry, it's not specifically to Agilent. The equipment manufacturing is not actually a manufacturing, is an assembly of individual components. Therefore, the factory, the production, the manufacturing site that does this assembly typically is not running at 80-90% capacity. In a case of a book-to-bill ratio greater than 1, it can be recovered pretty quickly from a manufacturing set-up point of view, if we stay between 1% to 10% more. If we are more than this percentage, typically, the bottleneck becomes the final testing. If we are, again, in a situation, 101-110 [sic], the impact is manageable because the testing capacity is already there. If we are in a situation more than that, then the testing capacity can be the bottleneck. It depends then on different set-ups in different factories and so on factories that can expand the testing capacity pretty easily with new bay additional or whatever and factories that do not have this possibility.

[[00:23:17]]
If the average duration from contract signing to true field installation expands past general and seasonal norms, at what point do customers begin cancelling orders to avoid delivery bottlenecks?

[[00:23:35]]
We are coming from the previous question where we said the book-to-bill ratio, if we are in plus-10%, it can be managed. Therefore, the impact to the customer is not significant, I would say, can be manageable. It's unlikely to have an order cancellation due to a delay of instrument delivery unless we are talking about government-type of situation. Excluding for one second, the government labs, it's very unlikely to have a cancellation. Now, the point of delivery times is more relevant before placing an order. The customer would appreciate the shorter timeframe or punctuality that you can offer, but once the order is placed, it's unlikely to have order cancellation.

[[00:25:18]]
Regarding the Ignite framework and broader margins, when Agilent experiences unexpected logistics disruptions or changing cross-border lane conditions, how many months or quarters of delay typically occur through internal product delivery timelines?

[[00:25:41]]
We are talking about logistic disruptions in this case. In, let's say, past experience, the recovery time out of logistic disruption has been in the range of three months. A case that is public, the Shanghai closure after COVID. This was unexpected and lasted for a long period of time. The recovery time has been in the range of three months. Considering the magnitude of this event and the recovery time of three months, we are talking about an event that has very big magnitude, unlikely to happen, but that was the timing. In other situations, like lane conditions and so on, the recovery time is shorter due to a reason to a fact that in this market, it's very unlikely to have a sea freight shipment. Typically, it's air freight. I would say 99% is air freight. In the case of air freight, you can reroute, let's say, easily, I would say, the material flow.

One thing to be added to this point, we are talking about supply logistics disruption, and typically, you have in mind a situation that happens in somewhere, whether Suez Canal or whatever. Actually, if we look back in the history, the major disruption didn't happen out of logistic disruption, happened because of an internal situation. Again, we are in a situation of public. Everybody knows that a company like Agilent or also other had to relocate three logistics centres in a period of two years. This was not due to logistic disruption. This was due to many reasons that affected or created the need of a relocation of distribution centres. That was the issue. In that case, there were known in the market situations that was not only a situation of not only, let's say, disruption in material flow for a period of time followed by a recovery, but a continuous, let me say, troublesome-type of situation for a number of months.

[[00:28:51]]
How directly does a tool manufacturer's on-time delivery metric influence its aggregate pricing power when negotiating large-scale asset replacement contracts with global labs?

[[00:29:10]]
There is the relationship between pricing power and larger-scale asset replacement as a result of on-time delivery, there is minimal correlation. There is minimal correlation that prompt good on-time delivery is not affecting the pricing, low correlation. Now, an erratic or, let me say, not in the market, on-time delivery can affect the decision of procurement of the platform, but once the decision is placed or let's say, it's not a driver of pricing. It's a driver for decision, but not pricing.

[[00:30:31]]
Could supply chain optimisations occur and expand fast enough to protect operating margins if there are immediate logistics pressures or inflation issues that occur for Agilent or other companies?

[[00:30:54]]
Inflation is something that's, here and there, inflation is everywhere. Of course, we are in a market that where inflation is affecting the cost of material is directly affecting the margins of the product. Therefore, yes, there are programmes to mitigate or improve. However, the major step of improvement comes from the new instrumentation, the new models that are launched in the market from the NPI flow. The more the company is able to design a new instruments, new models, considering the cost and being diligent on addressing the cost of the new instrument, the better is in the market. We are talking about instruments that are complex in nature, high mix, low volume, variability and so forth and instruments that are complex/in regulated market. There are restrictions on changing supplier, changing materials during the life of an instrument. It's possible. I'm not saying it's not possible, but the major source of cost reduction happens when the release of a new product or the design of the new product.

[[00:32:49]]
Are there any specific operational bottlenecks that dictate the timeline between implementing factory efficiency updates and realising visible margin expansion on the bottom-line?

[[00:33:04]]
These types of programmes, the time can be between one to three years. From the moment that there is a decision to launch a programme to the moment that you see, visibly realise the impact in the P&L, it's depending on the complexity of the programme, but can be 1-3 years.

[[00:33:33]]
What distribution bottlenecks typically occur across instrument fulfilment channels when a company compresses manufacturing overhead to optimise its factory margins?

[[00:33:45]]
If the, let me say, the reduction of manufacturing overhead is linked to management layers, reduction of layers of management, the short-term impact, typically, there is a value, but there is not typically consequences to having an engine that is not running anymore. The distribution bottleneck. If the manufacturing overhead that is affected, is in the range of manufacturing, engineering, or specialists in supply chain procurement, then there can be impact, but typically, the let's say, reorganisation in the short term starts with management.

[[00:34:59]]
Does implementing aggressive price optimisation metrics or concessions across a fulfilment network create customer retention friction with high-volume clinical contract lab accounts?

[[00:35:16]]
It depends on the definition of aggressive price optimisation. Are we talking about price increase, alignment of lease prices, or are we talking about actions to try to address the costs?

[[00:35:43]]
I was more so focused on price increases, but are there any other shifts? I was under the impression that price increases were the main consideration here, but if there were others that are more important to consider, happy to discuss that.

[[00:35:59]]
Just to make the pricing adjustment or increases, yes, can create friction. However, if you are talking about high-volume clinical contract, well, first of all, there are contracts that allows certain ranges of price fluctuations may be linked to indexes and so forth. If there is a contract in place, I mean, somebody might not be happy, but friction per se shouldn't happen. The second thing is that typically in this market and in large high-volume clinical-type of situation, there is, what I consider, a long period between notification and implementation. Typically, the period, commercially speaking, should be three months. Rule of thumb in the market. Actually, it's six months, which gives us the possibility with playing the order pattern, sorry, to extend the, let's say, the period of no price hit up to one year. If managed in this way, the friction, yes. Maybe somebody is not happy, but there is less friction. It's a way to mitigate the friction.

Remember that the effect of price increases or price variation, let me say, it's more with the smaller laboratories, 1-2 instruments-type of laboratories that place orders erratically consumables, not so much and so on, but that's actually where the price changes affect earlier the P&L. In case of clinical or larger-scale-type of fulfilment, it can be one year time.

[[00:38:20]]
Does an aggressive platform pricing layout build a commercial opening for lower-cost analytical tool peers to poach regional market share through aggressive promotional discounting?

[[00:38:35]]
There are differences by region. Typically, the North America region is the higher margin. Europe is in the middle, and China is the worst. That's how pricing is done considering many reasons with market considerations.

[[00:39:06]]
What operational complexities arise for commercial labs when migrating established testing methods from legacy set-ups to alternative gas layouts?

[[00:39:23]]
In a situation like that, there are two things to be considered. One is the, let's say, changes on the instrument readout of the instrument operations like sensitivity shift, linear speed shift and so on. These things when changes gas supply can be managed, can be evaluated and managed implemented in a relatively short period, say, 30 days. The, let's say, more complex situation is in the infrastructure. Imagine a change of gas and change of delivery of the gas, it means redo the infrastructure piping and so forth. That is the longer lead time item, but it's not the instrument. Another thing to be considered is the safety-type of situation. Is the new set-up considering new type of gases and with the safety regulation and so on, besides the real set-up of the infrastructure that is another long lead time item.

[[00:40:49]]
Global helium supply scarcities have forced labs to migrate their legacy chromatography workflows to hydrogen carrier configurations. How widespread is this technical carrier gas conversion trend within standard testing operations?

[[00:41:09]]
It's a very important question, a very understood question. I have just a little concern on this. Let me explain why. The use of helium as a carrier gas is predominantly on gas chromatography. While so far, we have spoken about liquid chromatography. In liquid chromatography, let's say, by definition, we are talking about liquids and not gas. Therefore, the helium-type of situation, there is little consequences to liquid chromatography. Therefore, LC and LC-MS-type of instrumentation. Now, if with chromatography, we are talking about gas chromatography, therefore, GC and GC-MS, that's a huge, let's say, operational issue currently, which again, I can expand if we want.

[[00:42:18]]
That would be helpful to dive deeper into that.

[[00:42:30]]
Helium is a gas that in the world, in our earth is becoming more rare. That's how the situation is. There is no, let me say, known mitigation at this point on the, let's say, extracting industry to increase the output of helium. That's structural. Therefore, the production of helium in a different method that is the extraction from the earth is not available at this point. It's not commercially available, let me say, at this point. The issue of the helium is an issue that it's hard to imagine a resolution in the next 3-5 years. It impacts the gas chromatography, not liquid. Now, changing from helium to hydrogen is possible, meaning that helium has a function of carrier gas, and hydrogen can have the same function of carrier gas. However, there are two things. One is, what we were talking in the question before about infrastructure set-up and so on. We were talking about safety. Hydrogen is flammable, highly flammable than helium that is inert from that standpoint. It's a major change from an infrastructure point of view. From an instrument point of view, it's a major change as well.

Let's say, at this point in time, I read the statistics that in that chromatography, 30% of new installations are hydrogen. There is already a shift to me is a high percentage, but that's what public statistics says. 30% of instruments, new installations going to hydrogen. To complicate even further, I said before the shortage of helium is coming from the extraction of the helium or limited in the extraction of the helium. Now, 35% of the helium produced in the world comes from the Gulf, the Emirates, Qatar area, which is affected in this moment by other problems that we know. The situation currently is deteriorated even further than the natural deterioration that I said before. Yes. I can talk till midnight about helium, but maybe I stop here if I've spoken too much.

[[00:45:39]]
Does the ongoing technical gas bottleneck act as a driver that shortens legacy hardware lifespans, or are labs extending the usage of older systems to avoid validation disruptions?

[[00:45:59]]
There are three possibilities to face this challenging situation. The possibility is to have a huge inventory, a big tank of gas, so that you have time to recover a supply chain shortage, but we know that that's probably not a viable solution. The second alternative is very strange and very funny in the sense that in the gas distribution system, it's very well known that there are inefficiencies, leakages in the distribution. Up to 30-40%. Companies or people do not really realise about that. I've seen in the market a situation of addressing this issue also by improving the gas delivery system inside the company and reducing the gas usage by 30-40%. We are in a situation of a shortage. Shortage that can be faced with the change of gas from helium to hydrogen, but another way to face this shortage is to reduce the consumption. Again, 30-40%, it's feasible.

This comes from two drivers. One, the leakages, the efficiency of the delivery to the point of use to the instrument. Second, the use of the instrument. It's like switching off the light when the light is not needed. Implementing something that distributes gas to the instrument in the moment that the instrument needed is also another lever to reduce the consumption. Consumption can be reduced pretty, I would say, easily with, let's say, reasonable investments, while change of technology, change of gas is, of course, a major upgrade.

[[00:48:38]]
Could a lab cleanly manage the gas configuration migration using simple on-site field conversion component kits, or does alternative gas use require a complete top-to-bottom tech refresh of the analytical hardware?

[[00:48:56]]
If the changes from helium to hydrogen, yes, basically, works really piping and safety standards and so on to be done. Second, the instrument, a gas chromatography can work with helium and hydrogen in theory, but in practice, you need a different system. There is work to be done also on the instrument itself. It's a major upgrade.

[[00:49:39]]
How effectively might gas chromatography upgrades incentivise biopharma accounts to proceed with fleet upgrades if they're facing multi-month method validation delays?

[[00:49:57]]
The only way that I can imagine is to go in parallel. The fact of switching from one gas to another, yes, it requires new, let's say, piping, new delivery, but it doesn't mean that you have to deinstall the old one. You can run in parallel with both gas delivery systems. By running in parallel, therefore, then you can have parallelisms, and you can schedule, it's possible to schedule things over time without doing a one-time major upgrade day one to day two. If possible, running parallel, but something needs to be done because, again, over the next 3-5 years, this shortage of helium will not resolve. It simply cannot be resolved, can go even worse.

I was just saying two methods. One is work on the efficiency of what you currently have and reduce 30-40%, maybe even strive to reduce 50% of the consumption, so mitigating the problem or change the technology, but change in technology requires more investment, more time and more people time and so on.

[[00:51:39]]
Given recent evolutions in the global semiconductor sector, how severely are localised microchip shortages or end-of-life component cycles disrupting the manufacturing lead times of flagship analytical instruments?

[[00:52:00]]
Thank you for this question because so far, we spoke about logistics, material flow and software, by the way, which is important. The hardware is important as well. If I had to say it's one of the biggest headaches, not only right now, but it started after COVID and never resolved. Actually, it got worse. In general, let me say, the pace of evolution in the chip manufacturing is very high. The pace of instruments, new product introduction of instrument is very slow comparing the two paces, first. Second, the culture in instrument manufacturer is not yet there. Consequence that we have seen in the market, and I'm not talking about Agilent, I'm talking in general, situation of new product introduction with the chips that are in yellow-type of situation. What does it mean yellow in the electronics world? It means that the chip is approaching the end of production rather than is end of production. Yellow is a warning that in six months or 12 months, whatever, the production will stop. Imagine situation of introduction of new product with chips that are already kind of approaching the end of production from a fab-type semiconductor producer. That's the situation.

Mitigation should be doing better scrubbing, better work, let me say, at scrubbing bill of material in the chips for new instrumentation, designing new PCA boards with possibility, compatibility and replacement possibility, which is more complex, but needed to be done. That's from a proactive point of view. Then, you have the installed base. Installed base, an instrument like an LC or GC can stay in the market. The use can be up to 10 years, let me say. Therefore, you need to have a possibility to support the instrument through the life cycle. For that, the way is, try to address by anticipating the need that you expect over the life of the instrument. Because if the chips are going out of production, there is no alternative rather than going through the second-hand market, distribution market and paying 10x more than the actual price.

[[00:55:43]]
How effectively could an automated fulfilment infrastructure, such as Project Ignite, isolate chip component scarcity before it impacts factory schedule attainment?

[[00:55:57]]
Project Ignite from the root of the project is a project to address the footprint to address the organisational model is not specific to address the chip semiconductor-type of situation. I don't see correlations between the two.

[[00:56:26]]
If component end-of-life bottlenecks push delivery lead times past historical seasonal norms, at what point do biopharma or advanced material accounts cancel their pending instrument contracts to source hardware from competing vendors?

[[00:56:45]]
We went through this before, meaning that the cancellation of an order out of a CAPEX, and it's an event that is unlikely to happen. Let me say, so far, the industry has been able to face this problem of shortage of chips with creative ideas, purchase on the, let's say, distribution market rather than the manufacturer and so on, so far, the situation has been managed. I think there is the possibility it will continue to be managed. However, not for infinite period of time, if the electronics on the new instrument will not follow design rules that are suitable for current electronics, then the problem will create what you said before, the cancellation. Customer will arrive to that point, so far, no.

[[00:58:14]]
When Agilent shifts its long-term strategic mix away from mature industrial hardware to higher-margin diagnostics and biopharma, what changes occur across its global distribution network?

[[00:58:35]]
Let me say that on the instrumentation point of view, the distribution model does not change a lot, let me say, instrumentation, instruments are built to order, make to order and so forth. There is no advantages to change the logistic distribution model simply because they are the same. Yes, you need to have a more, it's called the GLP good logistic practice in addition to GMP in the manufacturing set-up. You need something more, but you can implement it in the same, let me say, distribution model. Regarding consumable, that's a totally different story. Consumable and reagents, mostly reagents, the method of, let's say, shipping through global regional distribution model can be a bottleneck. How to implement a solution that is accepted by the market is to implement a direct shipment model from the factory to the end customer.

[[01:00:10]]
I'd like to close out by saying thank you for all your insights.

[[01:00:38]]
Thanks a lot.

Transcription ends at 01:00:43 of the recorded material