Seven Years. Five Weeks. One Million Dollars.

Announcer (00:04): Welcome to Raw Data with Rob Collie. We'll talk about AI and data for business impact. And now CEO and founder of P3 Adaptive, your host, Rob Collie.

Rob Collie (00:20): Hello friends. I have something exciting to report. We just saved a million dollars with AI, which is why it's been a couple of weeks since our last episode. We've been busy. Now when I say saving, no, there really isn't any million dollar line item on our P&L that's gone away. The only seven figure expense at our company is our amazing people. And we have not been using AI to reduce headcount. In fact, we've quietly started hiring again because, paradoxically, AI has increased demand for our team's time.

Rob Collie (00:54): But in a back office sense, we have realized about a million dollars in extra value because we just executed a complete rebuild of our website at p3adaptive.com. And we got a result that I can confidently say would have cost us at least a million dollars if we'd outsourced it, which means we got a result that would have been out of reach for us before. The version of our site we just retired dated back to 2019 and was largely unchanged since then. We did outsource a minor facelift in 2023, but otherwise it's been the same site for seven years. So until this week, our public facing website was like a time capsule. It described a groundbreaking company, but it was a groundbreaking company from a previous era. And honestly, it didn't even really describe the 2019 version of us all that well in hindsight. It never conveyed much of what made us different. Our philosophy, our personality, our approach, it was just kind of close enough even then

Rob Collie (01:56): in 2019. So why did we let it sit so long? Why did we address it now? Why do I say it would have cost a million plus if we'd outsourced it? And how did we do it ourselves? The answers to those questions are all intertwined with each other, but let's take them one at a time. First, why did we wait so long? Well, outsourcing website builds really, really sucks. Big, slow projects. Expensive. The previous site cost a lot. It was probably a quarter of a million dollars all in over time, once you include the minor facelifts after 2019, but it also cost us a lot of our own time up front, educating the external firm about our work, our industry, and our unique voice. And then when they came back with the new site, we spent even more time correcting their work, and it never got anywhere near perfect. Eventually, we just called it good enough out of exhaustion and out of a desire to cap costs.

Rob Collie (03:01): Kind of sounds a lot like BI used to be, doesn't it? Sink a bunch of your own time, then wait while they work in the dark for weeks or months. You hold your breath and wait, and then the big reveal, which is exciting for the first few seconds or minutes, but then once you look a little closer, the thing you're being delivered is kind of underwhelming. So then you argue and haggle for improvements, and you get a few of them. But then you eventually talk yourself into it being good enough so that you can move on to other things. No one is eager to repeat processes like that, which is why our site sat largely unchanged for years. And yes, ironically, that's exactly the broken BI workflow I wanted to fix when I started P3 Adaptive. If you go to our new about page and find the section called We Build in Plain Sight, you'll see how we actually work. Power BI was the disruptor that made our better approach possible all those years ago,

Rob Collie (04:00): and now AI allows us to take that same approach to building other things, too, like our own website. So why did we fix it now? Well, two reasons. One, we're so much more than just a data and BI firm now. We still do that kind of work, of course, but we've long since expanded beyond that perimeter. We've added custom AI and custom software to the menu of things we do for clients, and we've built quite a track record there over the past year. We've added a software engineering R&D team this year to help deliver on those new offerings, but also to start developing our own platforms, like this thing we call Faucet, that you'll see at the top of our homepage. Month after month, as those trains gained momentum, our website just sat there in its 2019 state.

Rob Collie (04:50): The distance between who we are and who the website says we are had become too extreme to ignore any longer. It wasn't just due for an update. It's long been overdue. So the first reason for why now is that AI transformation of our company necessitated a complete rethink of the website. The second reason is that AI transformation allowed us to do it ourselves. This website project literally started five weeks ago with our CTO taking the initiative over the weekend. No one asked him to do this to start a new site from scratch using Claude Code.

Rob Collie (05:30): He got us off the blank page in a big way with the first draft of an entire new site. He knew it wasn't going to be anything close to the final version, but he did also know it would get us moving. And it did. A handful of other people on our leadership team joined the project, myself included, and started iterating on it. We each brought something different to the equation. I, for example, remade the entire about us page and restructured the AI services page around the idea that most people don't even know where to start with AI.

Rob Collie (06:00): Our chief customer officer jumped in on many pages with the lens on clarifying our services and track record. And my wife and co-founder, Jocelyn, probably put in more time than anyone, bringing the design aesthetics up to a high standard and then constantly fixing drift every time one of us made a content change. With the help of Claude Code and GitHub, we got to build our website in the same way that we like to do BI projects build in plain sight. Everyone involved could constantly see the current state of the site and the latest changes and iterate and sometimes argue from there. For example, my first versions of changes to the AI and about pages were definitely in the right direction, but I got pushed back from others that I still hadn't really nailed either of them. And when I took those criticisms to heart and addressed them, the pages got much, much better. That's the sort of dynamic that played out over and over again over this five week process.

Rob Collie (07:00): When you outsource a project like this, you don't get to live that iterative constant improvement type of process that we just experienced. Seriously, go look at that build in plain sight diagram on our about page. This is how every technical project should work. Many touch points, many small check-ins with stakeholders, as opposed to one big ta-da moment at the end. And at every touch point, you can see it, shape it and improve it.

Rob Collie (07:30): That approach yields far better results than the ta-da approach. Always has. For BI, it wasn't possible before Power BI came along. And for websites, it wasn't possible until AI came along. In both cases, new technology made iteration faster. Now, let's circle back to that million dollar number. I'm confident we couldn't have gotten a site this good for anything less than that if we had outsourced it. We did price out an outsourced rebuild in 2024. We knew that we needed to do better than we had in the past, so we stepped up to a more premium tier of vendor. We did our homework, we picked an amazing firm, and the all-in cost was going to be something like $300,000. And then as we looked at the mockups, we were like, this is better, but it still isn't great. So we passed. That amount of time and money for good, but not great.

Rob Collie (08:30): Bad trade. And that's not an indictment of the firm in question. It's just the math of outsourcing something as custom as your own identity to people who don't even live in your industry. It's a translation problem more than a technical problem, and translation is slow and lossy. In our experience, getting from good to great through that process runs three to four times the original quote in time and in money. 300,000 times three or four is a million plus dollars. That's where the number comes from. And honestly, even at that price, you're not guaranteed to be thrilled. So our choice used to be outsource it or sit still. We chose sit still. Work was still coming in, keeping us plenty busy, and we've had other things absorbing our energy as well, like all of that inside-out reinvention of ourselves. Now, though, we have by far the best website we've ever had, even though today's site has a lot more work to do than the 2019 site did.

Rob Collie (09:30): It covers a broader array of interrelated services than ever and in the face of greater customer uncertainty than ever because AI is so new. I can't imagine trying to get this right with a vendor firm. This was literally our leadership team doing this. It was part time for each of us. I probably spent one full work week total on it as my contribution over the entire five week period. This cost us less time than it was.

Rob Collie (10:00): It would have cost us with a vendor, but it also cost us zero dollars in vendor invoices instead of a million. And then in the end, it's better than anything we could have expected, even from the million dollar process. And since we built it, we can keep iterating on it. We're going to discover that some pages don't work as well as we hoped, so we'll improve them. For a while, I suspect the site will change multiple times a week, even now that we're live.

Rob Collie (10:25): It's a brave new world for sure. I'm kind of blown away by how much we were able to do so well in such a short time. But I'm not sure we're going to add marketing website build-outs as a service we offer. This process worked because it literally was our leadership team building it. Like I said before, outsourcing your company's identity is inherently a translation problem more than a technical problem. I suspect that sooner or later, we will help a client overhaul their website, but our role likely will be more of a training and advisory on the process rather than critical path on the build itself.

Rob Collie (11:01): Zooming back out, the iterate in plain sight methodology that we founded our company around back in 2013 is now coming for basically every form of work that used to be waterfall style, long delay, and then ta-da. Just like in the post-Power BI version of the BI industry, many firms will stubbornly stick to the old style because it's all they know, and that style is very profitable for them despite being worse for their clients. But unlike the Power BI disruption, I'm not sure the firms who stick to the old style this time around are going to make it. It's not that the genie is more out of the bottle this time, so much as the genie is just so much more visibly out of the bottle.

Rob Collie (11:45): A lot of change ahead for a lot of industries. I'd encourage everyone listening to this to think about things that have seemed impossible to address in your business and think about them with a fresh mindset. Like your hardest problems, your biggest ambitions that have been out of reach. The scope of what's possible today is hard to even imagine. We, for instance, are at the cutting edge of this stuff, and yet we still discover, every day, something valuable we can now address for a client, or even internally, that we just as easily could have realized a couple of months ago. This site was one of those for us.

Rob Collie (12:22): Seven years of sitting still, and then five weeks to awesome. Anyway, go take a look at our new site p3Adaptive.com. And stay tuned for more news in our next episode, because there's a lot more change afoot, and I'm excited to tell you about it.