Should I quit my job to build this Monzo add-on?

Hey everyone!

I’ve been working on this: as a side project for a while now. Essentially it is an app that helps you discover cool new places like cafe’s, restaurants and bars by recommending ones similar to where you currently spend your money. You just log in and we automatically build up a profile of your likes and dislikes and then help you discover places you’ll love.

I need to work out if this is something worth investing more of my time to build, so all thoughts, criticism, idle chitchat is welcome.

Oh and if you like the idea, I’ve just launched signup for the beta, so make sure you reserve your spot :slightly_smiling_face:


1 Like

How are you monetising it?


I love the idea of this personally!

Always keen to find other recommendations to places, other than McDonald’s!


Haven’t decided yet, I was thinking it would be neat to link in to the OpenTable API or similar so that people could book a table through the app (and we’d take commission). Other options would be to run ads, although I am very conscious that these would still need to be relevant recommendations.

At the moment I just want to do it for fun and to help people discover cool places. Money comes second.

In which case I’d say the answer to your question is no, don’t quit your job until you’ve decided how this will pay the bills

1 bit of feedback tho, I think you need to be clear on your site about what you do with users data, how it’s stored etc. People want to know exactly what happens with their data and how to revoke permission etc
People need more than this

We take data security very seriously, that’s why we use the highest security standards


Awesome, thanks for the feedback

Sounds like a decent feature, I likely would only use it if it didn’t have any ads.

Nothing would make me remove the function more quickly if it were to show me ads, I’m afraid.

I’d also want to see how you use my data too.

If you can find a reasonable way to monetise it, great, but I wouldn’t take any steps until you can do so.

I will sign up for a beta though :slight_smile:

Think there’s your answer. Need to find a way to keep it going in your spare time, until you can at least be certain there’s money to be made in it.

At first glance it sounds similar to Forsquare, Yelp and the services built into Facebook and Google search. The challenge will be differentiating yourself from those services, particularly if you go down the ad-supported route.

Good luck with the beta though. :slight_smile:


I think that’s how it would be monetized realistically though. Places pay to be ‘featured’ higher up the list, Google-style.

Agreed, my frustration with a lot of these services though is that their rankings have a tendency toward mediocracy, and are not based on personal interest, just based on an aggregated popularity score.

How will your app NOT be based on aggregated popularity? Isn’t that how all ‘suggestion’ sites work? Or am I missing something?

So the app would take into account popularity, but popularity amongst people who like the things you do. The idea is that it learns about the type of places you like to spend your money and then helps you discover more.

For example, if a user transacts lots at sushi places, you could infer that they enjoy it, then recommending more sushi places they haven’t yet tried. You can give your recommendation a higher score than you’d normally see on Yelp or Foursquare because those aggregations include ratings from people who don’t have the same penchant for a bit of sushi :sushi:

Similarly, if your user rarely transacts at fast food places, you can discount any recommendations for those type of places, Foursquare, Yelp etc. won’t take your preference into account.

1 Like

But just because I eat sushi at restaurant A, doesn’t mean the sushi at restaurant B will be any good, so why would I try it? Not sure ‘hey this place here also sells sushi’ is enough to convince me to try something.

Yelp may not be great but at least it’s subjective.

Indeed, it was a slightly simplified example. You’re absolutely right that there are a whole load of factors you would try to incorporate; price, atmosphere, time of day, location, freshness…

I actually spent my MSc thesis researching algorithms for this and am planning on starting with something similar to what Netflix use for their recommendations. So in the same way that Netflix won’t just recommend any old movie of the same genre you’re known to enjoy, GoDorado wouldn’t recommend any old food from the same category you’re known to enjoy.

1 Like

Sounds intriguing. I’m still puzzled as to how you can NOT include user recommendations/ratings in such a recommendation service but good luck with it!!

I personally don’t like the idea of a computer trying to give me recommendations on where to eat, I prefer to get those from people that have eaten somewhere and experienced the food/services/atmosphere themselves.

This does sounds like a fun project but I wouldn’t quit my job because of it

1 Like

Sounds like an interesting idea indeed, however I’d chime in that asking some sort of rating (even just a 1-5 star) would seem to be essential to filter out the lower popularity places which are either really good or really bad. Also perhaps “How many people did you pay for?” would be useful to work out a per-head cost?

Also one other relatively easy but useful addition would be food hygiene ratings:

I think most people quit their jobs for this sort of thing when it has turned into an actual income OR know they can get it to that stage with a specific amount of time.

Good luck with this (I sincerely mean that), but a word of caution.

I have previously worked for one of the big online takeaway delivery companies in a team with a whole load of scientists trying to work out better recommendation algorithms based on the sort of stuff you mention using machine learning etc to spot patterns (I wasn’t doing the data science, but I worked closely with the guys who were).

Turns out, that even if you have data of millions of people placing millions of orders per month (week? Can’t remember the exact stats), it’s really hard to come up with an algorithm that works better than recommending a restaurant based on aggregated popularity (based on A/B testing).

I don’t want to dampen your spirits - maybe you’ll stumble on something they didn’t (and maybe they have something amazing now, this was a few years ago), just don’t quit your job just yet :slightly_smiling_face:


@VaticanUK, interesting! thanks for sharing your thoughts. I wonder what methods they’ve tried. One of the things that makes me think this is possible is that when I have been researching the algorithm, I’ve been quite successful in being able to predict missing places from a list of a user’s favourite bars/restaurants at a much better rate than predicting based purely on popularity.

@Gareth79 Thanks for your thoughts too! I’m really 50/50 on asking users questions I would want to make the UX as smooth as possible. Having said that, I quite like what Uber do when they just ask one or two quick Q’s on your last ride when you open the app, as it’s not too intrusive.

Interested to see if you mean this in a sense of ‘complete the list’ - I have three sushi restaurants listed, and you find two more - or in some way infer that of the five sushi restaurants in my area, I should replace one of my favourite three, with one of the ones I’ve not visited?