Matt Paice's One Percenters, Part 2: Revenue & Tech

Part 2: Revenue & Tech
Earning more and the tools that help

By Matt Paice

In Part 1, Matt looked at the cost and operations side of the one-percenter principle: tightening lots of small screws, half a turn each. Here he turns to the other half — nudging revenue up, and the modern tools that make it easier.

The Background:

Over a recent lunch at Quo Vadis, I was talking with Rav about our respective restaurants and how we try to turn a profit, and I mentioned to her the concept of “one-percenters.” The idea is: there isn’t one single lever you can pull to make your restaurant much more profitable, but there are a lot of screws that can each be tightened half a turn. Ten changes that each add one percentage point of profit become real money. So here, in no particular order, are some ideas for one-percenters in your business; they are not just cost reduction measures but also ways to tweak revenue up a little, or optimise your staffing levels. Most of them unfortunately require a certain amount of work; they’re not magic tricks. 

You may be practising all of these already. While some are things you do only once, others are worth re-doing a couple of times a year, so if nothing else let this article act as a nudge.

Kill your puppies

It’s common to calculate GP percentages for every dish, but that doesn’t give you the whole picture when you’re thinking about menu engineering. You can segment every dish on your menu into one of four categories: stars (high profit, high sales volumes), dogs (low profit, low volumes), workhorses (low profit, high volumes) and question marks (high profit, low volumes). ‘Profit’ here needs to be defined as high cash margin not GP%: a £28 dish at 70% is much more meaningful to your business than a £6 salad at 80%.

The point here is: remove the dogs from your menu, increase the price of workhorses and see if volumes hold, and then try to figure out with question marks whether the answer is to reduce the price or take them off the menu completely. And then think about pushing your stars’ prices even higher!

Price to expectation, not just GP

Some items are priced less by their food cost than by customer expectation. A side salad, Coke, coffee or portion of fries may support another pound or two simply because the market (or your competitors) have moved on. 

You can spend enormous energy protecting GP on expensive dishes while missing the fact that you’ve underpriced the smallest and most frequently ordered items on the menu. This is a constant review process.

Don’t let customers design your service

Airlines and hotels have long understood that not every unit of inventory has the same value and they rigorously practise what’s known as revenue management. Restaurants increasingly have access to the same tools: reservation systems can now shape demand as much as record it. I’ve seen a pattern in our bookings whereby there’s an increased share of reservations made within the last 72 hours. The right response is: hold back some primetime slots (eg 7.30pm) until the last three days to encourage those last minute bookers by offering them the most favourable slots. This is something of a trial-and-error process to drive bookings.

Cover pacing is worth a regular review: have you designed your reservations flow to maximise turns? It’s a moving target in the current economic climate so on some nights the answer might be to let the customers book whatever time they want, but for the most in-demand services you may want a pattern where 7-8pm isn’t bookable at all so as to force customers into turnable slots like 6.30pm and 8.30pm. I lose sleep thinking about how a misaligned cover pacing level could lose us two covers a week… which adds up to thousands in gross profit per year.

I’m seeing more and more restaurants inch towards revenue management techniques by offering variable pricing to fill inventory. Examples are free or cheap corkage on Mondays (Elliots, Hawksmoor), reduced prices Tuesday-Wednesday (The Clove Club) and à la carte menus available only after 9pm (BiBi).

A reservations system like SevenRooms has the ability to make offers available only to people (say, your database subscribers) with a specific web link which is useful if you want a limited number of customers taking it up, and don’t want to expose it to all potential guests. This is much cheaper than paying a third party to offer those discounts for you (apparently some restaurants are still on Groupon!).

Don’t optimise for the wrong risk

One missed table for two can easily represent £100 of gross profit. The problem is that while no-shows are painfully obvious, lost bookings are invisible. Cancellation fees should be about balancing the risk of a vacant table versus the risk of no booking at all in the first place.

A cancellation fee can recover some downside after a customer books; it certainly cannot create an additional booking in the first place and furthermore it can only deter some percentage of customers from booking in the first place. The longer your fee window and the higher your charge, the more effect this will have.

Airlines learned long ago that cancellation policies are really demand-shaping tools; flexible tickets cost more than non-refundable ones. The point here is: do you need a cancellation fee policy at the same level for every service? At Chishuru our cancellation fee window for dinner is just eight hours and that turns out to be enough – most people who cancel do so with at least 24 hours’ notice; and we don’t ask for card details at all for dinner bookings made with less than three hours’ notice. I see many restaurants that have dropped cancellation fees for certain quieter services completely.

Think about the customers who don’t book

At Chishuru, we have many many more people visiting our website than booking. We don’t know why but we can at least try and make it easier for guests: a ‘reserve now’ link on the home page, webpages optimised for mobile (two thirds of our website visitors are on mobile), and our Google listings are kept up to date.

If you haven’t set up your Google Analytics properly, that’s worth doing to see who’s sending visitors to your website. It may not tell you anything – the vast majority of our web traffic is people searching for our name direct – but there might be something interesting in there (for example, 35% of our web traffic comes from overseas, and half of that is from the US). Google Analytics (aka GA4) can be fiddly to set up, so using an AI agent inside your browser (Claude in Chrome, Chat GPT Atlas, Comet) can be useful to guide you.

Build your own tiny software

For the first time, non-technical operators can realistically build small internal tools themselves. Codex and Claude Code are surprisingly capable at creating simple web apps, and the hosting platforms they recommend (like Vercel and Supabase) are free at small scale. The sweet spot is repetitive, rules-based and operationally dull work – what drudge job do you spend an hour on each week?

At Chishuru, I built a linen-counting tool that lets staff enter counts, reads the consignment notes from London Linen, and compares our actual usage against what we are being charged for, so I can see on one screen how much we used on average for the last 13 weeks. Another operator I know built a tool that summarises his end-of-day reports (he likes them!) with AI into weekly and monthly operational summaries of the most important details.

The point is not that every restaurant should become a software company. It is that many small operational processes can move from annoying spreadsheet-and-WhatsApp combos into simple web automations, and give you time and energy to focus on revenue generation.

This also applies to number-crunching: where previously I did data analysis in Microsoft’s mind-bendingly complicated PowerBI app, now I can hand Claude our POS data and our SevenRooms feedback data and get a structured graph of customer feedback by table number.

Review your tech stack

Add up fees for POS, reservations, HR, inventory management, analytics, compliance and tipping software and you can easily be looking at £12,000 annually. These costs will fall significantly over the year or two as AI reduces the cost of software development. Hospitality software pricing still reflects a world in which building custom operational tools was expensive and technically difficult; that assumption is weakening rapidly. So keep your eyes peeled for cheaper software deals – this is how Blinq and Tebi are growing quickly in the POS market.

Even better: a lot of software we use isn’t just designed for hospo, it’s designed for everything from nail salons to retail, because that’s their business case. AI software development opens up the possibility of sophisticated tools aimed solely at independent restaurants, with the functionality and speed we need to make our lives easier.

With that in mind, I’m currently building a combined stock management, analytics and staff-scheduling platform that can come in under £75/month, with POS to follow as part of the bundle. I’ve built it for my own use first of all but if you’re interested in signing up for an app that understands that we build rotas thinking about “cl-opens” and “AFDs” and “ugh three doubles in a row”, drop me a line.

And finally… what else can you sell people?

This is really open-ended and will vary venue by venue, but what can you sell customers that isn’t already on your menu? At Chishuru we have an off-menu dessert that customers can order in advance for birthday people, and Rav is now doing the same at Gina with celebration cakes - in both cases using the SevenRooms ‘upgrade’ functionality. We sell jars of our house seasoning blend from the menu - at £8/jar it’s almost the definition of a one-percenter. FlatIron have previously sold their famous steak knives. Beyond the ubiquitous tote, what else could you sell to your best customers?