There’s a practice we have admired for a long time which we’re going to explore in this post to see if our assumptions are correct. Occasionally you see a new startup coffee business dive in the deep end and drop some serious coin on the best espresso machine they can find. If the cafe succeeds, this is usually a win because that location has had performance, modern styling and the latest technology on its side right from the beginning? My Grandfather would have advocated that approach. I can hear him saying “I’m too poor to buy cheap tools”. That’s all well and good if you’re talking about hammers and drills. But if you invest in the finest espresso machine right upfront before the business model has proven itself in your chosen location and things don’t go to plan, you will have dropped the best part of US$20,000 on a single piece of equipment. That’s nearly two year’s rent on a cafe in my town. So the alternative is to dip your toe in the water before you dive in. Maybe you install a nice old heat exchanger which you found on Market Place. Or better still, you just rent the time honoured symbol of ‘reliability + performance’ — the La Marzocco Linea Classic for a little while to see if things work out. 

Lloyd’s La Marzocco, Linea Classic (Photos Courtesy of Tortoise Espresso)

Old Lineas don’t have much that can go wrong on them. I have never come across one with a critical fault like a cracked or rusted boiler. But they’re not exactly big in the technology department either: There’s no PIDs; no flow restrictors; no insulation; no volumetrics (Some 90s Lineas have volumetrics, some don’t); no cool-touch steam wands; no flow profiling capabilities. But they’ve got neatly assembled chunky electrics; they’re made of quality stainless steel, and if something does go wrong, you’ll very easily find spare parts for them because they are ubiquitous in the coffee world.

You know I actually did buy an old heat exchanger on Market Place recently for a couple of hundred bucks. It was intended for my Sister-in-Law’s new cafe venture called Gornelly’s in Castlemaine. It actually looks super cool but it turns out it is missing quite a vital piece of pipe work with a very obscure gauge of thread on it which Gaggia themselves no longer stock. So this machine remains in my sister-in-law’s shed and instead of turning around and buying a new machine, Astrid Connelly and her husband David Gorniak (=Gorn-ellys) arranged to hire Lloyd Meadow’s old Linea Classic instead. 

You know Lloyd — he’s the anchor on our new lightspeed Espresso training course, Barista Zero. And he’s now the proud owner of a Synesso MVP — with volumetrics, pressure profiling, PIDs, flow restrictors, multiple boilers and pumps, and gorgeous styling. 

We asked Lloyd to walk us through the upsides of upgrading the machine at Tortoise Espresso. 

BH: In what ways has the MVP benefited the business since its arrival?

LM: We love having a bit more control over all our variables. It really helps when dialling in new coffee to have control over temperature and pressure profiling to really get the most out of them. We love having the option to manually profile a shot or have the machine do it all for us, makes a huge difference when we’re busy and just need to get really great results but a lot more efficiently.

BH: Knowing the success of Tortoise, if you could do it all over again, would you make the investment from the start, or sooner, or would you do it all the same?

LM: If I was doing it all again under similar circumstances and with a similar budget I would do it all the same. The  second hand market is flooded with really quality espresso machines, and knowing how similarly they scored; it just reinforces that a humble machine is more than capable of getting fantastic results. I would always prioritise getting a really good grinder as that’s the area that you’ll see a significant difference. Then as the business grows, your gear can grow with it, and you can easily make upgrades when your business can support it.

There is a 25-year technology gap between Lloyd’s Linea Classic (circa 1990), and the MVP (circa 2015). Notably, Synesso were pioneers of multi-stage pressure profiling; their machines are insulated, each independent boiler is PID controlled and the water passes through a flow restrictor before it enters each group. All these factors contribute to temperature stability and consistency between shots, and they trend towards higher extraction. Please note, we’re not looking to discuss the pros and cons of any single piece of this puzzle, such as pressure profiling. If that’s your jam, start with this lesson from our Advanced Espresso Course. Our objective here is to find out what gains (if any) Lloyd has gotten from upgrading from a 25 year old machine to an 8 year old machine. And to find out if Gornelly’s is leaving anything on the table, by going for the prudent, if somewhat conservative option of just renting an old machine with absolutely no automation at all. 

An Analogy for the Experiment

Lewis Hamilton on the podium

We designed an experiment to compare the performance both in terms of flavour and extraction efficiency. There are many differences between these two machines and isolating any single one of them would be almost impossible. Here’s a good way to think about how we designed this experiment to make it a fair comparison: Let’s say you ask Lewis Hamilton to jump in two F1 cars from rival manufacturers and compare their performance. Lewis will like each car to be set up in specific ways to help him get the best out of it. They might have different tyre pressure for instance or different amounts of suspension stiffness. But in the end, one car will go faster round the track than the other and people will consider that a fair test. The constant is Lewis and we trust his decisions on getting maximum performance out of each car. So we’re framing this post in the same way. Lloyd tuned each of these machines in such a way that he believes will achieve maximal performance and we turned up and scored each shot, and took TDS readings across a range of shot times from 14 seconds up 40 seconds. Our goal was to find the extraction ceiling and to identify the shot time band at which scores started to improve and then started to decline. 

Methodology

Mazzer ZM coffee grinder

We carried a Mazzer ZM grinder between each location to make sure this variable was removed from the equation. We also used the same Pesado HE% filter basket on both machines. Every shot was dosed with 18.5g of coffee and we attempted to yield as close to 45g out as the tech would allow. 

We used the same coffee from the same roast batch at both locations: An anaerobic natural processed catimor variety from Yunnan, roasted by Firebean in Castlemaine.

The pump pressure on the Linea was set to 9 bars.

Lloyd has the MVP set up to run a three bar ramp up to pressure; a 9-bar middle area and then the final 10g- of a 45g gram shot is ramped back down in pressure to 3 bars. 

Results

This plot for extraction percentage vs. shot time provides a clear view of the trends for both cafes:

Tortoise shows a steady increase in extraction percentage with shot time, suggesting that longer shots yield higher extraction initially, followed by a potential levelling off or slight decline at longer times.

Gornellys indicates a similar upward trend, though it appears to reach its peak extraction at slightly shorter shot times than Tortoise.

The regression lines (dotted) for each cafe help to visualise the relationship between shot time and extraction percentage:

Tortoise’s regression line slopes upward more noticeably, reinforcing the idea that extraction percentages increase with longer shot times up to a point.

Gornellys’ regression line is less steep, indicating a weaker relationship between shot time and extraction, suggesting more variability or inconsistency in their extraction process.

These regression lines visually confirm that Tortoise’s extractions are more predictable and potentially more optimised than those at Gornellys. ​

Sensory Scores

Here is a summary of the sensory scoring for each café, based on a scale from 0 to 6. 

We borrowed the zero to six scale from the WBC scoring protocol and aggregated the flavour, tactile and taste scores into a single ‘Overall’ score where 2 describes an ‘average’ espresso, three is ‘good’, and 4 is ‘very good’. The shots pulled on the Synesso MVP received higher sensory scores than the LM Linea on average by a bit over half a point across a range of shot times from 14 seconds with up to 40. 

1: LM Linea:
Average Score: 2.86
Score Distribution:
  • Common scores include 2.5 (3 counts) and 3.0 (3 counts).
  • Scores range from a minimum of 2.0 to a maximum of 4.0.

2: Synesso MVP:
Average Score: 3.44, slightly higher than Gornelly’s.
Score Distribution:
  • Scores are more varied, with 3.0, 3.5, and 4.0 each occurring at least twice.
  • Scores range from 2.5 to 4.5, showing slightly higher ratings on average.

This chart shows how sensory scores increase initially with shot time for Tortoise, reach an optimal range, and then begin to decline with longer shot times. The Linea at Gornelly’s has a flatter curve, indicating a less pronounced optimal range, with scores less affected by shot time changes.

Here is a scatter plot with regression lines showing the relationship between shot time and sensory score (WBC Score) for each location. The trend lines help illustrate how shot time correlates with sensory scores for both Gornelly’s Café and Tortoise Espresso, providing insight into whether longer or shorter shot times might be associated with higher sensory scores at each location. AT Gornelly’s the scores clearly trended downwards as shot times advanced — suggesting turbo shots were working better on the Linea than slow shots >30 seconds. 

Conclusion

If you run a statistical analysis* on the difference between the performance of both these machines in the way Lloyd chose to set them up, the sensory scores show a trend but not a statistically significant difference. But the difference in extraction efficiency on the other hand is very significant. 

The extraction analysis between the La Marzocco Linea Classic at Gornellys and the Synesso MVP at Tortoise shows a very low p-value of 0.00009 alongside a high t-statistic of 5.08. This extremely low p-value suggests that the difference in extraction between these machines is highly unlikely to be due to random chance.

So should you get a new machine and take the risk, or does this test make you feel a bit more comfortable that renting or buying a classic second hand machine is a better way to go? 

Here’s what David had to say on the subject: 

On the whole the machine doesn’t matter nearly as much as which beans you use. And customers definitely don’t care either. They only ever ask ‘whose coffee are you using?’ not ‘whats your machine?’ I can get a really decent coffee out of the old second hand La Marzocco. It would be interesting to see** the overall cost savings with higher extraction efficiency.

Both Lloyd and David seem to think that the stepping stone approach made a lot of sense for their businesses and I’m inclined to agree. Ironically, and in the interest of full disclosure, it’s worth noting that we were supposed to run this experiment a week earlier and the night before testing day, the solenoid on the Linea needed replacing, which sidelined the machine for a couple of days. I’d say that was bad luck. No matter how old or new your machine is, it’s always worth lining up a contingency option when your machine goes down because they all do occasionally — old or new. In this particular case, because it’s family, the contingency option was me loaning Gornelly’s the one group machine from my kitchen for a couple days while the Linea was in the pit lane. But she’s back up and racing now … to the boundaries of coffee.  

Does a more Efficient Machine Save You Money?

**In answer to David’s question, there is some potential saving in ingredient costs when you use a more efficient espresso machine. Across the shots we pulled, we measured an average extraction on the MVP of 19.77% at Tortoise and 18.52% at Gornellys. So with higher extraction efficiency, Tortoise could theoretically use about 93.6% of the coffee that Gornellys uses to achieve the same TDS. 

So if Gornellys used 1 kg of coffee per day (or 365 kg per year), and this coffee cost $50 AUD per kilogram, their annual cost would be 18,250 AUD. So based on our study, Tortoise  could make espresso of the same size and strength for just 17,082 AUD, leading to annual savings of around 1,168 AUD.

* In statistical tests, the t-statistic measures the size of the difference relative to the variation in data, and a higher t-statistic often points to a stronger difference. Meanwhile, the p-value indicates the probability that the observed results happened by chance; here, the tiny p-value highlights a highly significant difference, suggesting the Synesso MVP may consistently achieve higher extractions than the Linea Classic.