The New Year is nearly upon us, and if your organisation has been using the Google Search Appliance, by now, you know that 2018 is the year that the GSA is truly declared dead. And if you don’t have a GSA replacement in place yet, the urgency to find one may be keeping you awake at night (or at least, it should!)
Like many large organizations with their own IT department or development team, the debate about whether we should build something ourselves using open source, or use something “out of the box” has probably happened a few times, and may still be raging. We often get invited to join this discussion by customers and partners as they mull over the not so new idea of building their own enterprise search solution, but never more so than now, as the rush to replace the GSA takes place.
The idea of building something can be tempting, for a number of reasons. But it’s important to examine why this is appealing and also look at the reality of some of these assumptions.
Let’s be honest, this debate often happens because the price tag of an out of the box, best of breed enterprise search solution made someone in the company uncomfortable, leading to the question of “but couldn’t we just do this ourselves for less?”
The answer is, no, probably not. Building your own homegrown search solution based on open source components is rarely cheaper in the long run, and often carries a much higher total cost of ownership. And the initial spend is rarely as low as expected and leads to some uncomfortable discussions with the finance department when you have to return for unexpected funding.
Why is that? First, “free” open source technology is rarely free, especially not for large organizations. For it to scale, or offer the security, analytics and other critical capabilities for a larger company, it’s likely that you’ll need to upgrade to the paid enterprise versions. And often you’ll only discover these limits or requirements when you are deep in the project.
On top of that, you have to consider the cost of the development resources to build and maintain your solution. Even if they are internal and a “sunk cost”, the opportunity cost of using them to build a search solution must be considered. And we know, they are typically some of our more expensive resources.
With an out of the box solution, you’ll have a much clearer idea up-front of the true costs, and you’ll find a more predictable cost structure that will make your friends in the finance department much happier.
It will be easier.
No. It really won’t. The initial implementation is almost guaranteed to be more complex than you imagined, and as you progress you’ll discover hidden requirements or your users will add things they simply cannot live without. (Trust us – we’ve learned this first hand.) You are almost guaranteed to have scope creep and you’ll find your team scrambling to build capabilities beyond their expertise.
And then there’s ongoing maintenance. As you have no doubt discovered already, replacing your enterprise search is actually a pretty onerous project, so realistically you don’t want to have to do it again anytime soon.
In building your own custom search solution, you need to be ready to commit to years of maintenance and updates. And as search becomes more and more driven by AI and machine learning, it’s constantly evolving. This means you and your team will need to remain current and constantly adapt your in-house solution to take advantage of the latest advances. Your users will expect no less. As anyone who has ever lived with the legacy of a homegrown solution will tell you, this may be the most difficult aspect to manage (and the one you’ll regret the most as the solution starts to creak).
Which means you need to ask yourself, am I ready to commit my team to constantly supporting and developing this search solution, or can I see better ways to use their time over the next three to five years?
We’ll have better ownership.
By building your own custom search solution, you’ll absolutely be able to do whatever you want. But will that be the wisest course of action for the company? Are you certain that you have the right expertise in-house? Do you have data scientists and experts who are skilled in building the best possible search experience?
Will you be able to make use of machine learning and usage analytics to build a search solution that leverages the wisdom of crowds to constantly learn from your users’ queries and outcomes to deliver the most relevant results? Or, will you just end up with search that unifies or federates content, but provides no real relevance or recommendations as to the best content?
Real relevance is essential for your user experience. Find out more about how to understand the relevance maturity of your solution with the Coveo Relevance Maturity Model™.
And going it alone can be daunting. By using open source technology and building your own search solution, you’re likely to have very little, if any, support available to you. In fact, you may still require professional services or consulting (with associated fees) to accomplish what you want, and in the end, it may end up being wrong or inefficient, because it’s a one off vs being grounded in best practices.
And while every organization differs in the level of support they might need from “the experts,” it’s clear that taking advantage of the insight and customer success support that a best of breed provider can provide has significant benefits. Chances are, whatever issue you’re encountering, they’ve seen before and have a tried and true solution for you.
Plus, out of the box solutions have been honed based on years of experience, and many, many customer implementations, leading to continuous improvements to the search platform itself. In short, you’ll benefiting from all the projects before yours, with most of the common pitfalls avoided or addressed.
We’ll have the knowledge in-house.
The days of long tenure and lifelong employees are long over. If you build your own solution, the risk of critical knowledge and expertise residing in a few employees’ heads is high (because we all know how often documentation gets de-prioritized in a time sensitive project). If those employees leave, you’ll find yourself struggling with ongoing support and maintenance, or even to understand how the solution is architected.
And as the solution ages and maintenance and updates happen in conjunction with staff changes, you’ll find more and more of the knowledge about the solution gets distributed and lost.
In contrast, when you work with a search technology provider with years of experience, hundreds of client installs, and a system to safeguard knowledge about the solution, you mitigate the risk that comes with storing mission critical information with a few people.
We don’t need to make a long term commitment to a supplier.
In truth, commitment (or lack thereof) goes both ways. If you choose open source technology or an informal arrangement with a supplier, there’s no guarantee that the features and functionality you’re taking advantage of will continue to be supported or work the way you want. Essentially, you’ll have no real commercial relationship with the technology platform provider, and yet your business will be dependent on it.
While this arrangement might work fine for very small companies, it’s generally much less attractive for a large organization and increases risk significantly.
The risk is low – it’s just internal search.
There’s a lot more to enterprise search than you might think, and if you get it wrong, the stakes are higher than you might have initially imagined.
With poorly executed enterprise search you actually have the ability to negatively impact business outcomes within your organisation. Imagine internal users are searching for information that they need to make important business decisions or troubleshoot a customer problem. If search consistently points them to irrelevant results, or worse, outdated information, their effectiveness will be seriously compromised.
And as your users’ consumer search experience bleeds into their work environment, their expectations will continue to grow and the pressure on your team to offer exceptionally relevant search will increase.
You need a search solution that stays ahead of user needs, and is focused on delivering the most relevant results, always.
Making the case for out of the box search solutions
All in all, while it may be tempting to build an enterprise search solution yourself and hope for the best, it’s important to be aware of the risks and have a realistic view of the trade-offs you would be making.
The truth is that cheap search can be expensive. Without deep experience in building search solutions, a team that builds their own DIY solution is likely to fail to deliver a user experience that brings real business value or true relevance, without actually realizing the cost savings that drove them to build their own in the first place. Moreover, the ongoing maintenance and support will consume valuable internal resources over the life of the solution, while the risk of knowledge loss grows with every year and every staff change.
As the inhouse IT expert, your focus should be on finding the right best-of-breed solution that delivers relevant results and recommendations out of the box, that fits with your technology stack, offers lots of connectors to help you bring in data quickly and easily, and that allows you to efficiently manage the application over its lifetime. Moreover if you choose a solution that offers flexible user interface options and easy customization, you can still add your own flavor to the search experience for your users, without creating extra work for your team.
Ultimately, it comes down to understanding the impact and importance of good search on your organization, and making the choice that will deliver that today, and in the long term. Find out more about Coveo’s approach to GSA replacement on our Google Search Appliance Replacement page.