Founders of technology businesses typically find themselves having to outsource the job of building their product to what is known as a “software agency”. More often than not, founders turn to agencies that are based overseas, usually in Eastern Europe or Asia.
What is less well known is that, particularly first-time founders, often have an experience doing this that is fatally bad. It’s quite common for a founder to find, what in real terms is a huge amount of money, hand that over to an agency, and get nothing in return.
You tend not to hear about those people because after that crappy experience, they often give up their dreams of having their own technology business and go back to their day jobs.
Part of the problem is in how agencies typically manage their work, but – even something that would keep me awake at night as a CTO who’s been doing this for nearly 30 years – one challenge every type of business has is how do you go about qualifying an agency as being competent.
Like how do you know that they can actually do the job that you’re asking them to do? Let’s find out.
So in this video we’re interested in learning how to qualify an agency as being the right one for us. This means we need to take a long-list of agencies and whittle that down to first a shortlist, and then to the one that we’re actually going to give the commission to.
On my channel I have a lot of videos about how agencies work and how to get the best out of them, and I don’t want to repeat all of that here, but there’s a couple of things that we need to understand.
Most non-technology founders go down a journey of wanting to start their businesses and end up almost intuitively knowing that they need to find an agency and almost immediately on that journey they learn that agencies based in the UK are substantially more expensive than an agency overseas, and this is where they get tripped up.
(For a UK-centric audience I tend to present this as UK and “not-UK” – but for UK you can read any Western country, so any Five Eyes nation (US, Canada, Australia, and New Zealand) plus Western Europe.)
Outsourcing in UK and other Western countries are CONSULTATIVE in how they work, whereas any other type outsource – effectively an “offshore agency”, are STAFFING in how they work.
This means that if you engage a company in Bristol, Cardiff, Leeds, etc you will get a consultative-led engagement where the business seeks to inject technical leadership into you as a customer. This can be unbelievably useful, which I’ll come onto.
If you engage a company that is anywhere else in the world, you will get bodies thrown at a project. An offshore agency will build anything you ask them to build, without any real consideration to as whether you should or not. If you tell them you want the software equivalent of chocolate teapots that you’re looking to sell into nursing homes, they will just do that, despite the fact that’s the worst illustration that I could think of when writing this script.
Whilst all outsourcing costs are fundamentally based in the cost of labour – that difference between consultation-based and staffing-based approaches is something you need to factor into the price.
A UK agency will cost you £80/hour compared to £50/hour in Eastern Europe (e.g. Poland) and £20/hour in Asia (e.g. India). New businesses with no in-house technical staff should ALWAYS choose the £50/hour Eastern European route, but that £30/hour price differential you should notionally think of buying advice.
The reason for this is that startups are operating in full-on experimentation mode, because very new businesses have not yet found their product-market fit. Agencies are, frankly, terrible at dealing with clients who are experimenting – they exist practically solely to provide staff to existing technical teams in established businesses, and established businesses are never LOOKING FOR product-market fit because the reason why they are established.
That difference – spending £30k more to outsource to the UK – the value of that consultancy could save more than just your blushes.
The point of putting this in early doors in this video is that these are very different beasts, and you can’t compare like-to-like. You either need to decide to offshore and not invest in the value from the consultancy-led approach (which is fine, but I’d rather you knew you were doing that), or you do want to invest in that consultancy (which is also fine). If you want to do both (which is also, also fine) the process we’re about to go through – you have to do twice.
The safest way to find an agency is to ask someone that you trust to give you a recommendation. We know this is generally the safest way to find someone to help us no matter what we’re doing, e.g. if we need someone to help train the family dog, or fix the roof, or whatever.
I will qualify this though in that it’s very important in this instance that the person you ask has direct experience of the agencies work. I’ve been bitten by this when seeking certain types of professional services – just because someone “knows someone”, it’s not enough, they need to have some ability to qualify the person they are recommending.
The next trick is to socialise the query on LinkedIn. This can be a bit hit and miss as it depends on how widely LinkedIn broadcasts your post looking for support, so you may have to do this a few times.
If you’re going down the road of trying to find an agency in the UK, there is a trick you can do with Google which is to run a query in a private browsing window like “director custom software uk linkedin” and Google will surface a fairly enormous list of options.
However, unlike searching on Google normally where you get the best results at the top, that list from Google will be completely random. What you can do is scroll down until you get a hundred or so names, copy and paste them into ChatGPT prefixed with: “extract the names from this list as a table:
Then put that table in Excel and choose the ones you fancy – basically from that very long-list you might as well just stick pins in the list whilst blindfolded until you get 10 that you want to take on for further qualification.
Annoyingly, that list only seems to work when operating Google from within whatever country you currently are in – for example modifying that query to find agencies in Poland or Warsaw or wherever doesn’t return results as effectively.
So how do you find agencies in other countries?
Honestly, I would not try and find them proactively. Realistically, you are looking for 3-4 agencies to give you a quote, and if this is your first time doing it and if you don’t have experience at commissioning software projects from agencies, you best line of defence is personal recommendations. I would just keep digging and asking until you found the ones you wanted to ask.
I would also not trust any agency who approached you using a cold direct message on LinkedIn. It seems really harsh to say this, but the reply rate to cold outreach on LinkedIn is so low that for every 3,000-4,000 people they message, they only get 1 reply back. Finding a software agency is supposed to be a proper partnership, and the agency needs to be able to see if they will be able to work with you as well. If they are only getting 0.25% reply rates on their marketing, they will desperately grab at the chance to work with you – i.e. they will not qualify you at all, they will just listen to what you ask for and present a proposal.
There is a similar issue with organic search results on Google. The agencies that show up when you search will be those that have consistently spent £500 a month for at least 24 months to end up in those top listings. Google search doesn’t work for software agency listings because the only method to appear on those listings is consistent month-in-month-out application of spend either with an SEO agency, or with Google AdWords pay-per-click. If you are relying on those Google listings, do the same trick that I did before – scroll to get a few hundred results, paste them into ChatGPT, as it to create a table, and then stick pins at random into the listings to get ones you might want to qualify further.
Just to orientate you into the problem, the core thing you are trying to do is to spend a LIFE-CHANGING some of money with a company that YOU DON’T KNOW. You either are pre-seed investment in which case you’ve managed to find something like £50k to spend on the project, or you have got your seed investment, in which case you could easily be looking to spend £125k-£150k on that project.
Although the objective is not to screw that up, fundamentally you need a backstop – and the backstop you get is being able to take the supplier to court if everything does go to custard.
Herein lies your first problem. If you are based in the UK and you commission an agency also in the UK, and you need to litigate, you can do that. If you commission an overseas agency, as a tiny business who’s just scraped together the cash to pay for the project in the first place, you effectively have no legal remedy.
So back to my earlier point that a £50k project in Poland may cost £80k in the UK, that £30k difference is not just consultancy; there’s a backstop in there whereby you can take legal action.
(As a sidebar, I am not a lawyer, and I would strongly advise doing everything in your power to avoid litigating a complex IT project, but this is a factor that is almost universally not appreciated by first time founders.)
There are two reasons why size is important and why it comes first in our qualification process. Firstly, they need to be big enough to do the job well, but not so big that your small project gets lost. Secondly, if you do end up in a dispute with them, they need to be able to actually pay you back the disputed amount should it come to it.
This is another situation where there is a big difference between the way UK and non-UK agencies operate. In the UK, if you ask how many employees a company has, the answer you get back will (generally) be the number of PAYE employees on their books. So, if you ask a UK agency how many developers they have, and they say “30”, it’s quite likely they have 30 UK resident, PAYE employees. This is down to a sort of cultural convention – part of the unwritten rules about how companies within the UK negotiate business with each other, but also partly because a UK agency knows that what you are paying for is this consultancy idea. The number of “bodies” the UK agency has isn’t as important, because the value is in the leadership and judgement.
If you are in the UK, you don’t even need to ask them. If they’re the right size business, you can just look for them on Endole. Their “number of employees” metric when you look can be a bit sketchy, but if they’re a tiny business the number of employees won’t show up at all.
It’s very common with offshore agencies that the relationship they have with their developers is looser; oftentimes is setup more as a loose connection of self-employed associates. The agency may say they have “200 developers”, but what they mean is they have the email addresses of 200 developers. I am perhaps being uncharitable, but these businesses exist to solve staffing problems and it’s well worth understand that that fact colours how they present their businesses to the world.
As a result it can be hard to properly size an offshore agency, so you have to go more directly.
What you are looking for is an agency that can run around 5-8 projects of your size at the same time. Notionally you can allocate about five people to a development project of the type that you are looking to do, so as a result you are looking for about 25-40 development staff, plus a handful of admin and management staff.
Because an agency will not be expecting this question, once they know what your project is, you can “cheat” by asking them, “how much projects like mine can you run at the same time?”
The caveat is that they need to actually be at a stage where they could start to put together a price proposal, but we’ll come back to that – but once you are there, if they answer comes back at about “5-8” for a UK agency, you’ve sized it right. For an offshore agency, because of the different staffing model, you’re listening for a response that’s about four times that – although if they give you the 5-8 range you should accept it.
Anything smaller than five projects like yours (including yours, I guess), and the agency is too small for a business LIKE YOURS. If you are a technology person, you can go with smaller agencies. If you are an established business, you can risk it if you’re confident you can litigate a remedy if needed.
Anything bigger than eight projects and you won’t get the attention that you need.
So to reiterate, coming into this section you need to have your long-list of agencies that you want to work with and you’re looking to whittle that down based on size. Now, we can look at capability.
Counterintuitively, the capability of the team is less important than the points we have looked at above – i.e. we’re looking to reduce bias in how the long-list of suppliers is created, and the using size of the organisation as an initial proxy for capability on the assumption that a business that has grown for to be a reasonable size ipso facto has to be able to deliver. What we’re looking for next is whether they are right for YOU.
There are two things that are NOT necessary when it comes to the capabilities of the agency. The first is their preferred technology solution – within limits, the actual technology that an agency wants to use on your project won’t matter. This wasn’t the case 20 years ago, but in this modern era, most difference in technical approaches come down to preference. (There are some exceptions, but as I say generally you don’t need to stress about this.)
The second is whether the agency has done a project similar to your one. It stands to reason that this would be a good signal of being able to achieve success in your space, but the reality is that even similar projects with similar goals have much deeper differences than you might thing.
However, both of these factors when you reverse them do become relevant. Most technology businesses that are founded by non-technology people are B2B “software as a service”-type businesses. There are fewer startups that focus on B2C. o– but B2C has some odd specificities in terms of how the product has to work that means it is essential that if you are building a B2C product, you have to find an agency with specific experience in B2C. Likewise with B2B, but seeing as most solutions are B2B, it’s generally not hard to find an agency with B2B experience.
Similarly if you are building a solution that’s fancy, such as D2C – direct to consumer, or even B2G – business to government, you should find a specialised agency. I would also suggest that for this weirder business types, as a startup you absolutely should not offshore those, and you should definitely prefer to find a consultative, domestic outsourcer because those types of solutions get quite sprawling from an overall IT systems strategy, and you will need specialist help with that.
In a similar vein, some industries have a preference for technical approaches, and you may not be aware of that at the early stages of your business. For example, the NHS in the UK has a strong preference for Microsoft-based technologies because Microsoft has had their notional tanks parked on the lawn outside of the lawn of a notional NHS HQ for 30 years, and there’s an unconscious bias there. Other industries have other biases, so try and discover those early on and align with that.
Now that we’re at this point we’ve trimmed our list down by using recommendations and apply randomness to remove bias, qualified each by organisation size, and then looked at alignment of capability to our business model, we can look at how we start to interact with the agencies.
You may be thinking, “wait Matt, haven’t we been interacting with the agencies until now?” and I will say, “maybe, but it wasn’t serious”. Now it’s time for serious discussions with the agency.
To preface this, one thing to be acutely aware of is that agencies are desperate for your business. Finding customers for bespoke software projects like MVP development is very, very difficult. As a result, the agency will be eager to do this project, even if they know deep down that they may struggle to deliver. You have to be the sensible one and do all the qualification.
What you ultimately need is to get 2-3 of the agencies to produce a price proposal for you, based on the specification of what you are trying to build. This is where things do get difficult because you won’t have a specification, you will have to synthesise one, somehow.
A specification is, essentially, just a list of functions and behaviours that the software will have when it’s finished. However, it’s a specialist technical document as a non-technology founder it’s very unlikely that you will create a spec that is canonical enough to quote against – and my earlier point stands in that any agency may be so keen to land the project that they will produce a quote on the assumption it’ll be roughly right and they’ll work it out later. (It won’t be roughly right, and they won’t be able to work it out later, and all the fallout from that will come back to you and your wallet.)
The most effective way to synthesise a specification is to ask one of the agencies to undertake a scoping exercise where they undertake a short consultation exercise with you to produce that specification. You will have to pay for this, and I would budget about £4k to get a UK agency to do this and a £2.5k to an offshore agency.
What you will get from a scoping exercise is two documents – one being the specification, which would have been produced after extensive interviews with you, and a proposal containing a price. The price is literally created by going through and counting up in detail a time estimate against each of the things the specification describes. That time estimate is then multiplied by a notional hourly rate, and the total given.
A soft advantage of the scoping exercise is that you’ll know whether you like and trust the staff at the agency that is working with you. This is clearly a very important indicator – do they get what you are trying to do, do you think they can help you, and is there an alignment of values.
A corollary disadvantage is that you will tend to select the vendor that did the scoping exercise, i.e. the actual process of commissioning someone to do the scope can by default select the supplier for the final product. This isn’t really what you want to do, so my advice here is to take the specification produced from the scoping exercise and send it to one or two other agencies on the list and see what prices they come back with, and also when you do that try and engage when them as much as you can to get a feel for whether you like and trust them too. Then at least you’re making a decision based on kissing three frogs, rather than just the one.
This is about as far as you can go as a non-technology person. The next step if you are a deep technologist is to apply some stress tests, i.e. to actually look to actively gauge their technical nous and delivery capability. But you would need either a technical co-founder, or a Fractional CTO, or a best bud who was a senior technologist to do that. In the absence such a tame person at your beck and call, the things we’ve been through in this video will stand you in good stead to find the agency that’s right for you.
SIGNOFF