Search Information Manager: Solving the Unit of Measurement Nightmare
With the release of HawkSearch Search Information Manager, on June 17th, Michael Benedict, the HawkSearch General Manager and Jonathan Meyer, HawkSearch's Senior Solutions Engineer held a presentation on optimizing your customer’s experience. This webinar demonstrated how HawkSearch has finally solved the nightmare of Unit of Measurement conversion with the HawkSearch Search Information Manager.
Unit of Measure Challenges and so much more
Have you ever wondered why search results vary based on similar search terms such as why ‘6 inches’, ‘6 in’ and ‘half-a-foot’ don’t return the same results? Does this still happen whether your site’s product data is originating from a PIM, ERP or even a simple spreadsheet? There has never been an adequate solution to remedy the problems caused by unsolved data issues that reveal themselves during the search process until now.
Normalizing data may sound simple, but in terms of search, it is actually quite complex. You have to account for not only how the product data is stored, but also the terminology your website visitors use search. Conventional solutions like synonyms cannot solve issues like fraction detection (½ = .5 or ¾ = .75), symbols of measurement (1 inch vs. 1” vs. 1in), measurement phrasing (6 inches vs. half a foot), part number inconsistencies (dashes vs. no dashes), and so on… During this webinar, Michael and Jonathan explained:
•How data in search terms becomes normalized
•Where SIM augments PIM, ERP, and other systems to cleanse data problems
•How to optimize the search engine indexing stage for maximizing results
Watch the full HawkSearch Search Information Manager Webinar below:
Why HawkSearch Search Information Manager?
When looking to close the gap between your PIM/ERP and your visitors, HawkSearch SIM provides groundbreaking new functionality. This feature, now a part of the HawkSearch platform (below) helps further clean, enrich, and connect the data you have about your products or content to how the customer is searching for them.
So, what does HawkSearch SIM solve?
There are many problems with marrying data and how visitors search for information. According to Hubspot, most compelling websites make it easy to find what the visitor is looking for. This is true with ecommerce sites as well. Information such as exact product (SKU), products that solve a problem (i.e. wall-mount bike rack), information (best gaming computer), answers to a question (how do I make a return?) and why to subscribe or buy your product (membership details) are examples of some of the most important tasks your website needs to accomplish. At the end of the day, websites need to connect your visitors with the right products.
Navigation vs Search
When visitors land on your site, there are only two ways they find what they are looking for. Navigation, often by choosing a category and then selecting filters via the navigation bar and search, typically by directly searching for a product by typing into the search field and engaging facets. But these are often supported by two different technologies pieced together in the wrong place. Raw data contains specific sets of statements that typically do not correspond to what the visitor is looking for.
Some examples could be as follows:
1.ProdA | 6in | W
2.ProdB | 5” (text) | D
3.ProdC | 2 ft | L
4.ProdD | 2x4x8 | LxWxD
Cleaned and enriched data allows for the above information to be normalized, allowing for the client to obtain the proper, desired results. This would require manipulating the data request into a response that the visitor is looking to achieve. For example:
1.ProdA | 6 | in | W
2.ProdB | 5 | in | Depth
3.ProdC | 2 | ft | L
4.ProdD | 2 | in | L | 4 | in | W | 8 | Ft | D
Once the data is normalized, then search semantics can be applied to produce the desired results as shown here:
Solving the Unsolvable
When connecting visitors to products on your site, converting data must be completed in the search engine to produce the desired results. The normalized data and the sematic query processing would need to be combined to provide an intelligent response for the visitor’s request. With HawkSearch SIM , connecting the product data and converting the data occurs concurrently, removing the discrepancies that would return an inaccurate response.
HawkSearch Search Information Manager combines the semantic analyzers and the data analyzers into the SIM artificial intelligence core producing a SIM index database. This database then can be used to produce the desired result from the visitor’s request.
How to Use HawkSearch SIM
There are many use cases that were discussed on the webinar. Some of the most popular practical use cases we have found most effective with our clients are as follows:
•Unit of Measure and Value Conversions - 6” = 6 in = ½ ft = 6 inch = 15.24 CM
•Fraction Detection – basic conversions (½” = .5); how people measure (8/16 =1/2 = .5)
•SKU Search - stripping special characters, mistaking letters with numbers, removing blank spaces, treating as a phrase (for just the SKU field)
•Synonym Patterns - detecting patterns and building synonyms (volt, -volt, v, watt, -watt, w)
•Phrased UOM Detection - Split Token Detection (6inx4inx4in = 6in x 4in x 4in), phrase detection (“6in x 4in x 4in”), UOM Phrase Flip
•Data Enrichment - combining attributes to generate a new attribute, extensible
•Parts of Speech Synonyms - building synonym tool, using WordNets to augment synonyms, industry-specific accelerators for synonyms, handling copyright
•Text Summarizations - long descriptions, policy summaries, PDF & document search
•System of Analyzers - drives customers, to the correct products no matter what variation of the search terms they type in and what the underlying data is like
The More You Know
If there is anything you would like to hear about on future webinars, feel free to email us at firstname.lastname@example.org. Want to learn more? Listen to the replay of the webinar and contact email@example.com to learn more about how HawkSearch Search Information Manager can help your business.