Go big or go home: The Sterling Archer story
Sterling is a US based data company that provides real info on prospects for companies and political fundraising whom in collaboration with we designed multiple digital tools & products to help their workflows and processes.
The Case
In the USA the political system allows for direct lobbying of individual citizens, in particular with requests for donations to support a party or candidate for office - not just of overt politicians, but for other public servants such as law enforcement officials. To some extent this sort of activity also goes on in other parts of the world, but in America it has become formalized, and recognized as a legitimate way of raising interest and support in the form of sometimes very large amounts of money. Over time this system has also been adopted by non-profit organizations wishing to raise consciousness and funds, and until the very recent past, has meant vast call centers full of operators telephoning supporters and putting political candidates in touch with them. How successful those calls were depended to a considerable degree on how accurate the data was about potential donors. Was their name, number, address, voting preference and other crucial information actually recorded correctly in the available databases? The answer to that question was frequently No, which meant that the armies of call center workers were wasting huge amounts of time chasing down accurate information, or even failing to connect with prospects at all.
Little better than random
Enter Martin Kurucz, CEO of Sterling AI, and his co-founders Josh Lorah and Caroline Jaros, who around 2020 were all working in political campaigning in the USA - specifically for a congressional candidate. Martin explains, “It's like a huge sales operation where each candidate is trying to sell themselves to donors in the hope of getting contributions. We were working out of a Contact Management operation that was specifically built for candidates to call individual donors, and there was a score on the screen for each of the donor profiles. My background is in research and data science, so I was always wondering about that number - from 0 to 100 - with 100 being most-likely to donate.” The problem, as Martin saw it, was that the scores for successful donations looked suspiciously little better than random chance, so he ran a study on his own time to check out his intuition. And yes, he proved his hunch, and took his study to the boss of the center. “I said, look, my background is that I can build algorithms, so let me figure out a way to prioritize these donors, because our congressman has only so many hours in the day to contact them.”
Changing the pitch
Using his algorithm, a month later the campaign was raising more than twice as much money as before. Martin and his future partners saw that they were on to something, and with this particular campaign over and job done, realized that other politicians could benefit from the technology. With CRM duties parked for the time being, Martin was out of paying work, and returned to finish his college studies. Josh meanwhile was applying for other jobs on other campaigns, and midway through one interview, saw the light and completely changed his pitch. It wasn’t for himself that he wanted to get another grind-it-out job, but rather for the algorithm-data-thing that had proved so successful for the congressman. There was one small detail that Josh neglected to mention - no company actually existed - but his pitch was convincing enough to get a new campaign boss interested, at which point some very intense founder’s work had to be rapidly undertaken.
Becoming the largest Data Broker
Martin explains that, while we might think of ‘data driven’ companies as being highly technical and analytic, “In reality they were just a bunch of people with Google Sheets trying to scramble around for data, with no actual idea of how data works. We saw that we had something that could make it significantly easier for them to reach out and raise a whole lot more money. We're good at logistics too, so we built a CRM staffing operation. And basically, that led the way for us to become the biggest Data Broker in the country, because very quickly we learned that our data is better than everybody else's. We make it so much easier for campaigns to actually find the right donors and make the whole process of getting donations so much more streamlined, easier and faster, in terms of how it's going to be scaled.” Indeed by 2022 Forbes magazine was crediting Sterling with changing the way the midterm elections were shaped, predicting that Artificial Intelligence, ‘Will become a permanent feature of Democracy.’
The biggest Data Broker in the country - just like that, in less than four years. Presumably there was a huge amount of work involved, but how did it all happen so swiftly? “Basically, the short story of why Sterling was so successful is because we understood one fundamental truth about all outbound communications like texting, email, whatever. That truth is: it's one thing to find and filter the perfect target you're trying to reach. It's also really nice if you can find the perfect message for them. But that is worth virtually nothing if you are either missing their contact information or you do have contact information, but it's wrong.” Martin says the Sterling database contains some 30 million political and nonprofit donors but the company had to make sure that as many as possible had the right information about them, which meant doing a lot of fixing. Some sources of commercially available data were actually only about 50% accurate. So, ‘A ton of time and money’ was invested in the first eighteen months of the fledgling company’s life to refine the contact data engine, which started to return between 20-30% more conversions overall. That’s on average, so some conversion rates were even more impressive.
What’s in a name?
OK, so Martin, Caroline and Josh had a burgeoning success on their hands, which had grown from almost nothing. At what point did the name Sterling come into the picture, and is there a story behind that? As Martin tells it, he was keen on finding a ‘British-sounding’ name, and most people assume that it refers to the pound sterling, or sterling silver. Not so. It’s actually a reference to a favorite TV show, Archer - an animated series featuring the adventures of ‘a suave master spy, Sterling Archer.’ It seems that, as with the fortuitous pitch made by Josh Lorah, the path to naming the successful business wasn’t overthought. A good name, OK, go with it! Which now explains the second string to the Sterling bow - pun intended - the subsidiary company, Archer.
With Sterling getting such traction in the Data Driven Acquisition market in the USA for political campaigns and charities, the attention of businesses was also soon turning in its direction. Could the same techniques be applied to sales and b2b operations? Well yes, for sure. Josh Lorah happened to be on a call with ‘one of the most aggressive salespeople you’ll ever meet’, just checking out the possibilities of a credit line, should it ever be needed. So while being pitched to by the aggressive salesperson, Josh upended the flow, and identified the lack of data-driven acquisition for the salesperson’s own organisation. “Basically he ended up doing what he’d done for the original Sterling pitch,” Martin recalls. “Josh pitched that we have Archer for b2b, when we had nothing even close to Archer at that point. It was one of those moments, where once again we realized: OK, people get this, people want this thing.”
Adding the AI dimension
Getting into the b2b space has hugely widened the potential of the Sterling / Archer axis, because now the offering isn’t just limited to politics and non-profits, and is certainly not limited to only the USA. Once again Martin and team built a whole new platform which he says is still in the early stages, but enables more and more accurate contact data for both consumers and business professionals alike. And, he says, it does this better than any other platform that is out there. “You can not only feed it a list of people and find good contact info for them, but clients can supply a list of people they already have contact info for, which we know is potentially bad. And we can find the bad ones and replace them with the correct data. It’s a pure value-add.” He points out that for a big company in the b2b space, especially in a ‘high ticket area’ every potential lead is valuable and could be worth hundreds of thousands of dollars. It seems that Archer is on target.
Which is where the AI component comes in presumably? Martin recalls how when Sterling began, the partners had a ‘decent’ advanced statistical method for searching out donors. However after half a year in the space, the company partnered with a specialist AI operator. As a result Sterling now has a considerable reputation for being an early adopter of AI technologies, and Martin points out that while much of the world sees AI as being generative - such as ChatGPT - the deep learning and data science part where Sterling and Archer operate is five to ten years ahead of the current curve. Sterling was already doing very well, but when AI entered the frame to do the hyper-targeting of potential donors, previous methods were quickly beaten by around 2x.
The perfect puzzle piece
Before the advent of AI in the business, Sterling was a big employer of people with its CRM part of the offering. This came at the time of the pandemic when suddenly it was accepted that a company could be staffed by remote-workers, so there was no further need for large call centers. Gradually however the personnel-heavy roster has decreased, with now core teams of about twenty people in each of the Sterling and Archer operations. Which means that partners are needed to fulfil many of the applications and developments that are ongoing. Partnering with Ergomania Digital Product Design seemed to come with similar ease and a sense of serendipity as with many other key decisions for Sterling. “I knew András,” [Dr. Andás Rung, CEO of Ergomania] says Martin. “We go way back and he is actually one of the best friends of my oldest sibling.” In the earliest days of Sterling - before it was even a named operation - Martin and András happened to meet up, and Martin outlined his ideas for the new company. András quickly recognised the business potential, and suggested ideas about a front end, which Martin tucked away for future use. When the time came, he contacted Ergomania to explore building the CRM platform for political outreach. The Digital Design company already had a lot of experience with call center-like elements, CRM, and payment processing, mainly gained through banking and Fintech clients. “I knew that András was really good at what he does, and the trust was already there, so I asked him to pitch to me and my co-founders. We saw then how these guys are real experts.”
From early on Ergomania team leader and Head of Design, Balázs Tárkány-Kovács was heavily involved, and Martin says of him, “I really loved working with Balazs. We just clicked, he's so incredibly good at what he does. He's invested in your project, and he really wants to understand it - he's the kind of guy who does not leave the job at the office. He's literally thinking about all these ideas all the time. I can bounce something off him at 7pm. I can send him something, and he's like, ‘Oh, that's a great idea. I was just thinking about that, just seconds ago.’ So, we clicked, basically as I was putting together my little army. I knew that for design, they were the perfect puzzle piece.”
Design in two directions
Checking in with Balázs reveals a man who is very circumspect about client confidentiality, and who is careful not to say too much about projects which are only just being launched into the world. However he does give an example of searching and filtering, as to how Archer could operate. “It could be very simple: I need carpenters in New York. But then I go to another level because I only want people between 20 and 40. Or I want to look for healthcare professionals in Kentucky. You can do this with logical operators like and/or, and so on. It can be quite simple, or quite complicated.” Balázs describes the design philosophy as always needing to be in two directions: Firstly for simple everyday use where someone who doesn’t understand databases, statistics or demographics can still easily get accurate results. The second approach is to input and search in a far more sophisticated and complex way, using both the logical operators, and complex mathematical expressions. So Ergomania’s initial input was to tailor designs for both types of end-user.
Enriching the data
Balázs points to the adoption of AI as being significant, and very useful to the essential process of query building, “You can express yourself in common language - with words - and the system will really understand what you want.” The system can also rephrase what you're telling it so that it acts like a ‘reality check’, prompting further refinements. OK, carpenters aged 20 to 40 are needed in New York, but do you want furniture-making carpenters or building-site carpenters? The Archer offering is continually expanding, but Balázs says that with the core public application, the user states their needs though language or mathematical expressions, and then the system tells them what kind of data it has, and at what cost.
That’s the buying part, but there’s also enrichment , where a client can upload a database which Archer then checks to assess if this can be enriched. Balázs gives the simple example of a database with 200 records, which on checking can be revealed by the system that of those, the email address is missing or incorrect for 180 of the records. Can all of the missing email addresses be found? Maybe not, but perhaps (in this example) 100 correct email addresses can be added. Let’s say that the system ‘knows’ that Ergomania email addresses are firstname.surname@ergomania.eu, then it’s a relatively simple operation to match up known team members with their correct email address.
Managing the whole lifecycle
That’s now, but what about the original work Ergomania did for Sterling? Balázs describes it not just as data handling, but also as a Contact Management System which deals with the whole lifecycle of a call, starting with an sms text, then a voice call (which can also be recorded). Messaging by text, voicemail or email can be sent out while the call is in progress, or after. The system then has a record of when the call took place, and whether it was completed, or had to be parked for any reason. If so, did the person called commit to a callback date and time? - Future appointments can be created, with handovers to other call center colleagues if the original operator is not available. “So you have different campaign managers, and different agents who are calling people for the campaign, and so on. And the original system was capable of handling all these levels,” says Balázs. “From the ‘do it yourself level’ - for a local Sheriff's campaign for example - through to a really complex large team level. At that level, you need a quite complicated user management system.” The system also scrapes all sorts of interesting information from social media and websites which can be aggregated to provide ever more touch points with the person being called. The operator can know about family, pets, hobbies, educational record, and - importantly - donation record. In the USA for instance, all donations made are a matter of public record. So rather than cold-calling just a number and a name, there is already a highly personalised set of rich data available to the operator - warm calling, as it were. “On the front end,” Balázs expands. “There are different surfaces where you can manage these connections, and this scraping and these APIs.”
The fractional approach
OK, so it’s clearly a complex, multi-faceted system which has been brought to fruition in very short order. Sterling / Archer and Ergomania managed this with regular structured weekly meetings and workshops of three to four hours at a time. These were initially in the Discover category of collecting information and research data, as well as studying business strategies. The second type of meetings was around the Define phase - what the use cases were, what the aims were, and how to structure the data. As Balázs says, “At the beginning, we are dealing with raw data, or maybe presentations. Later, we drew mind maps to some wireframes, along with other diagrams which described the layout of the system.
We basically express ourselves in a wireframe. So if we start to design something - and by design, I don't mean graphic design - but just how it functions. In the beginning we express it to basic, low level wireframes: Not too much detail. We express how we would solve a problem. And then if everybody agrees on those basics, we go into the details week by week and month by month. And basically, that's what happened.” Balázs also points out that even when a system is implemented, it’s rarely the ‘final’ version. As soon as real users get onboard it’s necessary to listen carefully to what they’re saying, and then make adjustments accordingly. “So the design never ends with the first version. It really just starts there.”
Discovering immersed in the political culture
“Balazs told me that this will be a huge project and he explained to me the dynamics of how these political campaigns work,” says Regina Eszes, Senior UI-UX Designer at Ergomania - although when she was first involved with Sterling she was a new incoming member of the team. The scope of the work was immediately appealing to Regina, not least because she has always had an interest in the USA, and politics. To help prepare she immersed herself (in her own time) in docuseries on the American political system, and in particular the fundraising process. “It was really surprising to me that they are not only calling really rich people to fundraise, but ordinary people who might just donate fifty dollars.”
This initial investigation was part of the Discovery phase, which in time moves into the Define phase, before landing on the Design phase. The Sterling team mapped out some of the features that they wanted to see on their platform, and from there Ergomania began with a Visioning workshop which examined the purpose, ways to fulfil the purpose, and critical questions relating to the intended initiative. The Discovery phase also involved interviewing users, including political candidates, call time managers and the Sterling core team, usually from a questionnaire of some 10 to 15 questions. Best practice research was also undertaken, where Sterling provided a list of their closest competitors, which Ergomania then investigated. In some cases only short demo videos of these platforms were available as they are only open for the intended audience, but the team were able to compile a lot of information on the competitive landscape. The upshot of all this activity was presented in Discovery workshops where the presentation of issues were grouped by task, ranked by severity, and underlying reasons and possible solutions were mapped. This helped pinpoint a host of issues, particularly those gained from contact and interviews with users.
Defining the user journey
Now that there was increasing clarity, the project began to move into the Define phase. At this point the work was still being referred to as Arrakis, from which Sterling would eventually be born. Discovery and Define phases were initially kept separate from Design as there were many factors in the first two parts that could massively change the scope of the project. The Define phase included feature workshops, user journey workshops, and the work of creating personas in a handful of Persona workshops. Regina says that high level persona creation can sometimes take years and is an ongoing iterative process. However in this instance a lower-level method was chosen in order to empathise with Sterling and always keep their needs in mind when designing. As a result the teams worked quite quickly to create personas, and then later merged some of these. The personas were based around the call time manager, different types of donor (such as the 50-dollar donor, and the high-value donor), and the different candidates. These could range from a nationally active politician with a large support team, through to an individual running for local office, who could be operating the system solo. When Ergomania creates a persona, it’s like a person, with a name, age and occupation. “All the basics that can be used as a summary of them,” Regina explains. “Then we go into details, and usually define their goals, their motivation, their needs, their frustrations, maybe some information about their knowledge and their technology skills.” Creating personas helps define the user journey, and the system had to function at multiple levels to enable different user journeys, and allow for call handling, calendar functions, and secure payment collection.
“There were a lot of challenges in the planning,” Regina recalls, “particularly in the area of callbacks. That was a really challenging task for us to figure out because sometimes people weren't available, or they’d say ‘Call me back tomorrow,’ or the next day, or whatever. So we had to manage all this.” The overall process was very iterative, and also encompassed built-in limitations, such as if six calls were made without a successful donation forthcoming, the contacts would be dropped, so as not to bug potential donors any further.
Designing for the multiplicity of interconnected parts
Regina reflects that, “Balázs mentioned that this project was one that really stood out for him, not just because of the complexity, but also because we could use almost all the methods that are available for UX people. With a lot of clients, we cherry pick two, or maybe four tools, but here we had much more variety and choice. What I’m always interested in is having a project that has multiple modules to it, where there are many connections between the different parts. And I like to see those connections, understand them, and also understand that if we change something in one part, how will it affect the whole.” The Design phase encompassed defining modules, low level wireframing, mid-to-high level wireframing, look&feel UI design, design system creation, and design of all the screens. For Sterling, the complexity and interdependency of the many parts included how the call time managers and candidates needed to log calls as they progressed, recording what was being pledged or actually donated at the time, along with being able to follow up and check those donations later. Regina says that just the screens created for that activity alone went through ten iterations before the final format was delivered. And then, after nearly four months of the Design phase, the project parameters changed as the Sterling team decided to reframe their offering. A hair-tearing moment? Not at all - Regina says it was a business decision that made sense and resulted in a far better platform, which went on to become the Archer offering.
Flexing with the business decisions
These business decisions happened two or three times overall where radical changes were made to refine what Sterling was bringing to life. This included dropping some of the features that supported multiple logins - where the political candidate, calltime managers and others could be live on a call. Instead, the focus became more on solo-candidates and smaller teams that could handle being on the calls at the same time, but in other ways. The original working title of the Arrakis project then morphed into the spinoff titled Archer, which in turn became the main product. Says Martin Kurucz, “We spent a good year building out the designs, but we didn't end up actually making it for business reasons, but from design to dev we were very happy with all the work Ergomania did on the project. Archer was already potentially happening and we knew that it was one of those projects where if it takes off it's a rocket ship. The addressable market is so big, that if this works, it really works.” Martin also compliments the Ergomania team - and his own - for being able to deal with working on such a ‘fractional basis’ which provided the ability to scale up or scale down as needed, and to do so nimbly. The advantage of that approach was ‘huge’. And as for working with Europeans and time zone differences, Martin makes light of it, saying that he feels European design can be really good, especially with tech, and that transatlantic working is not necessarily about price points.
The nerd at the whiteboard
So what does the future hold for Martin and his twin companies? The answer is a little different for different parts of the business. For Sterling, the main political data operation, there's nowhere much higher to go. It’s a model tailored specifically to the American political system, and is already number one in the market. In case you hadn’t noticed, 2024 is a presidential election year, but the following years will not be, so there are also cyclical limits to growth. As Martin says, “We're not planning a ton of new products on the political-slash-nonprofit side.” On the other hand, Archer is getting traction. “Our main focus right now is to tap into the sales intelligence base. That's Archer. So the big bet right now for the next 5-10 years is Go big or go home. If we can do this, if we can scale up, and do what we did with Sterling, there's 5 to 10 years of work that I need to deal with, to scale that across the world. So that's the big bet right now. If that doesn't end up working out I'll have to go back to the drawing board. But we're very young and - even though we had the big success and got a ton of recognition for it - we're still very hungry. There’s this weird little industry where we made a big success, so if there's a time to be incredibly ambitious, it’s probably now.”
Martin is a runner, footballer, tennis player, traveller, and has ‘a ton of other hobbies’, but also claims to be something of a nerd, and is never happier than when he’s in front of a whiteboard figuring out algorithms. He started off interested in politics, and then had his passion somewhat dimmed by working in the field, but he does still get excited about solving tough intellectual challenges. “These big intellectual problems, throw them at me, I love working on them. And if I figure them out, that's incredibly satisfying. That’s the thing that makes this work fun, and I know it's cliché, but I wouldn't call this a job because a job is a duty. This is fun, and I feel good about that.”
Want to read more
about UX, fintech and banking?
Subscribe to our biweekly newsletter and get the latest of our news and thoughts!