Top 8 Evaluation Criteria to find the right Analytics Solution
TOP 8 EVALUATION CRITERIA FOR SEARCH AND AI DRIVEN ANALYTICS
We are seeing huge amounts of disruption in the Retail sector with institutions like Debenhams and TopShop disappearing from the high street, largely due to Covid-19. We have also seen some great success stories with ECommerce retailers such as Wool Couture, Depop and Papier whose profits grew during the past year. What’s clear is that the retail companies coming out the other end unscathed or even in a better situation are those that have kept pace with innovation and digital transformation. At the heart of this, is data!
The big question – how can you get more people in your
business, making better data driven decisions?
1. Search Experience
Gartner predict that by 2021, over 50% of analytical queries will be generated via search, natural language processing, voice, or automatically generated. But not all search is the same, so it is important to understand how each solution works. For example, some only search pre-built reports and dashboards, some only look at metadata, some merely return a list of matches, some provide a single answer or even worse a long list of ranked search results of pre-built reports that the user has to wade through. Meanwhile, the newest breed of search-driven analytics engines search through all the underlying raw data, compute results, and then present charts and numbers based on those real-time calculations presented through best fit visualisations, for that particular question.
2. Search Intelligence
Most businesses run on many different data sources, most users don’t understand how all this data relates to each other. All this complexity should be kept away from the user. A smart, search-driven analytics platform should deliver an experience that recognises patterns, has spell check, understands synonyms and offers suggestions as they type based on other users activity – such as how Google’s predictive ‘type-ahead’ feature works.
3. Chart Creation
For a search driven-analytics product to be effective and therefore adopted widely throughout the business, it must remove as many barriers as possible. This isn’t the case with traditional or legacy BI tools where the wait time between the users query and the visualised result is often slow. On average it takes 5 days but we speak to some customers who say it can take weeks and even months before the business user gets the data they are looking for. Instead, a user should ask a question, just like they do for example when asking Google ‘what is the weather tomorrow?’, and the product itself should automatically create the dashboard on the fly – there should be no wait time. The user shouldn’t have to choose which type of chart they would like the numbers to present, the solution should present them with the best fit visualisation based on the question that was asked, but give the user an option to select different chart types if needed. It’s reported that 23% of current BI users are comfortable creating charts & graphs. A good analytics or BI solution should democratise access to data, remove this complexity and enable the least technical users to ask questions and get on with their day.
4. AI Driven Insights
What if your analytics solution could answer the questions you didn’t even think to ask? What if this intelligent machine could access huge data sets, generate thousands of questions across billions of data points and find hidden insights, outliers, anomalies, trends in your data. What then if it could surface these insights to you as a user, because it knows the type of person you are, it understands what’s important to you and it know’s when these insights should be delivered. What if you could work with this machine, and help it to learn from you?
5.0 Speed & Performance
The power of Google is achieved because of its ability to search every website across the whole of the web. However some BI tools, often fall short of allowing users to analyse data across the whole of the business with some restricted to files on your local machine. IT and data teams are often left with having to make decisions around which data sets to include, as they know that reports can take overnight to run because of problems with scale and performance.
6. Total Cost of Ownership/ROI
Implementation Costs – what are the costs to deploy the product into your environment. A good BI solution should work straight out of the box and you should be able to easily connect to your different data sources.
Implementation time – is your analytics strategy tied into a new product launch. Time to value should be considered and in some cases it could be best to focus on specific and targeted use cases.
Infrastructure Costs – where will the technology be deployed, what hardware will be needed, what about storage, maintenance and support etc. Consider working with a partner that can offer a fully managed service, so you can scale up and down on demand.
Resource – how many technical staff will be needed to administer the solution. As we’ve mentioned earlier in the article, try and find a solution that offers true self service for the end user, without the need for a data team to keep churning out reports and dashboards.
7. Security and Governance
Most data & BI solutions have enterprise grade security and governance built in these days and can integrate into your existing directory services through protocols like LDAP. Do the search box and search results feed into your access rights and rules, so users only see what they have access to see?
8. Training & Support
BI adoption today sits somewhere between 20% and 25% in an enterprise and many traditional & legacy BI solutions require lengthy training courses, for so called ‘Power Users’. The average duration of a beginner BI training class is 3 Days.