A discussion of the FAST search engine which is now part of SharePoint Portal Server. We explore the business opportunities in both employee-facing and customer-facing scenarios and discuss the differences in the technology from the perspective of readers that are already familiar with SharePoint.
Home » What We're Thinking » FAST SharePoint Search
I am a runner.
As a reflex, when I think of who I am, I think of myself first in the context of when I am at my best. In my mind, that is when I am out for a run.
The Boston marathon is 1 month today. I am consumed with excitement.
I have no lofty expectations for myself with respect to performance. I am not Kenyan. But I am a competitor at heart.
In running, when you are Alison Davies not Paula Jane Radcliffe, the competition is in your head. A Personal Best is the most I can hope for. So how do I prepare mentally and physically to give myself the greatest chance of achieving that PB?
I search of course.
I search for training regiments, route advice, accommodation recommendations, nutrition suggestions, weather patterns; I even search for people. I search for people who can validate or discredit the information I uncover, based on their experience having read the same articles, reviewed the same videos, followed the same diets and most importantly run the same race.
Even though my searches begin in the bath tub - flipping through magazines, while soaking in Epsom salts, surrounded by bubbles and basking in the afterglow of another long run - they are inevitably taken online after the water cools.
So here is a recent frustration I encountered in my search for Boston preparation material:
Last week I was given a copy of iRun magazine when I purchased a pair of shoes at the Running Room. Unlike Runner’s World, which feels like a companion piece to Groundhog Day, (seriously, I am convinced they have a database of 100 articles that have been recycled at random and put out in rotating issues for the last 10 years!), each issue of iRun feels unique, relevant and fresh.
After reading the issue cover to cover, I decided iRun was my favourite running / fitness magazine. On Monday, between meetings, I went online. I was confident the iRun website would be the Holy Grail of Boston insight I had been looking for.
Look closely, or go to the website yourself – www.irun.ca . What is wrong with this picture?
No search. Not poor search, no search. I could not believe it! My instant reaction to this busy, albeit beautiful, homepage was, ‘I don’t have time for this’. And I was gone.
Fortunately for iRun, I am feisty. They couldn’t get rid of me so easily. Returning to the site on Tuesday, I realized that someone, with enough patience, interest and sheer determination, could sift through the navigation layers and content silos to find snippets of what s/he might be looking for.
Funny though, as much as I dislike Runner’s World, where do you think I resumed the hunt for Boston material last night? Mind you, I had to be feisty there too. Sure there was a site search, but it was no Fast search, I tell you that!
For the love of God, can't somebody build me a Boston Marathon Topic Page??? What’s a Topic Page you ask? Give me a call: 416 619 1122. I can talk Topic Pages for days.
Better yet, let me tell you my ideas for the Boston Marathon Topic Page ... I can see it now and damn, is it ever sweet!
Recently I have found myself in client meetings raising the topic of search as a strategy. A critical distinction I have to make, time and time again, is that search as a strategy does not equate to search as a product. As I reflect on these conversations, I cannot help being reminded of a former colleague who I hold in the highest esteem. He was the first to impress upon me the importance of asking ‘why?’.
Why? why? why? Never lose sight of 'why?'.
· Can Fast support federated search? Yes.
· Can Fast accommodate faceted search? Yes.
· Can I use Fast to search multimedia files? Yes.
· Does Fast search structured and unstructured data? Yes.
· Will Fast return search results in less than a second? Yes.
And on, and on, and on...
9 times out of 10, the answer is yes. When it comes to features and functionality in the context of a search platform, there is a reason Gartner ranked Microsoft Fast Search #1 in the category of information access technology.
Is everything you can imagine available out of the box? No. Through customization can it be done? Probably, yes.
So let’s save ourselves the time of running through a features checklist shall we?
Rather, let’s take a step back and refocus on what is really important: the strategy behind the search. What are your business objectives for the site?
1. Increase revenue?
2. Reduce costs?
Whatever they are – the business objectives, should guide the solution discussion. The solution discussion should not be guided by the features of a product.
· Why is faceted search relevant to your users?
· Why is searching multimedia critical to meeting your business objectives?
I often think back to many years ago when I was writing proposals at a furious pace. The organization I represented at the time had developed an application development platform upon which all of our websites were built. Every time we built a feature we looked for the opportunity to reuse it elsewhere. We had a polling module, a simple search, survey template, calendar tool... oh and we were hot on Flash too. Flash just for the sake of Flash. Flash website intros? Yes please, I’ll take two. **
What did this mean back then? It meant that 90% of my website proposals were reduced to a features and functionality pitch. 'You can have this, and this, and this – oh yeah, and this and this. All that for this price!'
Rarely did I stop and think about why a polling module would enhance the user experience; or more importantly why a polling module would serve my client’s business objectives for the website.
The concept of why is so simple: If you cannot answer why the feature serves to support the business objectives then it is not only unnecessary, but possibly a distraction from the business objectives.
Fast may be 'the best there is' when it comes to search platform technology. But so what? What does that mean for you, for your business?
Before we compare and contrast the features and functionality of one product versus another, we need to understand why the features and functionality even matter.
** We were not the only interactive shop guilty of this mindset 9 years ago. And like our industry peers, we matured in our proposal approach over the course of the 2000's! **
Is your head spinning trying to figure out the different search options from Microsoft? With six different choices available today, I can relate. How do you know how they are different? I’ve pulled together this diagram to help.
The following describes how search functionality expands as you go up the product line:
Two products are planned for the Office 14 timeframe that will bring ESP fully into the Microsoft product suite: FAST Search for Internet Business for external users, and FAST Search for SharePoint for internal users.
We’ve talked in earlier posts about search-as-a-strategy instead of as-a-feature. In the feature world, there are several low-cost options. Install a Google Search Appliance. Fire up a Microsoft Search Server. Both options can deploy with an extensive audience for under $30,000. Box checked.
With an increasing shift towards search-as-a-strategy, however, there isn’t a box to check. Instead, an exercise is required to identify tangible business pain that a line-of-business system can address. It is not unfamiliar territory – what is new is the idea that search-based solutions can and must fall into that category as well.
Consider the example of getting procedure information to front-line retail banking staff. Timely and accurate information increases customer satisfaction because the job gets done right and quickly. It reduces costs, because employees do not need to contact internal call centres to have their questions answered, and because it reduces rework in the back office. This is a search problem at its core, and one that FAST is very well suited for. However, the solution is more than $30,000. How do you get started?
An envisioning process determines a business case (see our process for customer facing scenarios). The business opportunity needs to be defined and quantified, the solution outlined at a high-level and costs determined, and an understanding for how the solution will fit into your business and taken to market is required. From there, a business case is defined. That puts tangible go/no-go decision information into your hands.
How do you know if a search solution is even worthy of that exercise? Consider some back-of-the-napkin math to validate the opportunity. Looking back at the branch staff scenario, here is an example. Consider the following assumptions:
If we only factor in reducing internal call centre size (hard cost savings) and value time liberation at cost, what kind of result does this drive? Here is a summary of the outcome:
The three year opportunity varies from $9,000,000 to $21,000,000 depending on the size. Targeting roughly a 10x ROI, that puts the all-in investment range required between 1 and 2 million. Certainly enough room to confidently start an envisioning exercise to nail down the real numbers. This process leaves out the additional benefits of back office savings, additional efficiencies, customer satisfaction, and employee retention. Here is a link to the spreadsheet, if you are interested.
A posting by Kara Swisher reveals a sneak peak at Kumo, a potential update to the Live Search experience. Part of the posting includes three sample screen shots of the beta search experience.
I want to highlight a key difference: the left-hand side of the mockups now include a FAST-like navigation experience.
The categories appear to be dynamic, and for a sample search of ‘audi s8’ we see parts, used, forum, accessories, sale, images and video.
An interesting move looming, and it highlights the opportunity for our customers. If the web-based search experience is trending in that direction, do you want to be ahead of or behind that trend?