In our recent quality review, it came to our attention that many analysts have been doing many mistakes that we consider now to be common issues in your market.
Please take time to review this document as it will help you improve your quality.
If anything is unclear or you need further support, please don’t hesitate to contact us on our mapping portal https://homeworker.custhelp.com and we will get back to you.
The Autocomplete feature is intended to save the user’s time by presenting relevant content as the user is typing. Each task present us with the ‘query string’ initiated by the user and the result in form of Autocomplete ‘Suggestions’.
As analyst our task is to:
Rating Relevance
How to Approach Initial Relevance Rating?
Note: Distance and Prominence will be next in Order to determine the final rating.
Considerations | Autocomplete |
Identifying the intent | •“Is the query string complete or incomplete?” •“Is the query misspelled?” •“Would I search for the suggested result by typing this query string?” •“How likely is the user be looking for this suggestion given the query and the location intent?” |
Primary Intent | A suggestion that is useful to the user holds significant prominence, or directly satisfies a possible category intent of the query string. There can be more than one primary user intent suggestion per query string. → Assign initial rating of Excellent (User Intent) |
Secondary Intent | A suggestion that is not as prominent as other potential suggestions or satisfies the intent of a possible category or query string to a lesser degree. → Assign initial rating of Good(User Intent) |
Unlikely User Intent | A suggestion which matches the query but is very unlikely to be useful to the user. → Assign Initial rating of Acceptable (User Intent) |
Non -Relevant Intent | There are issues that make the suggestion useless for the user, such as no match to the query string. → Bad (User Intent) |
Type of Items we will be reviewing | Address, Business/POI suggestions, Query suggestions and Category Suggestions. |
Location Intent | •Explicit: The query indicates a specific location or area where suggestions are expected •Implicit: The location expectation is not given in the query. So we must use context clues in form of User location and Viewport to determine where the suggestions are expected. |
How to establish Relevance | 1. Matching Suggestion 2.User Intent 3.Distance/Prominence |
Name and Address Verification | Please review Country Specific Guidelines for Specific Requirement for Autocomplete Tasks. Utilize official resources to confirm data. |
General Match 4.1.1
Query String | Possible Suggestion | Explanation |
Mg ro | Mahatma Gandhi Road, Bengaluru Karnataka | Match: Suggestion matches the query string and is prominent suggestion. |
Starbucks MG Road Ground Floor, Kids Kemp Mall, Trinity Circle, Mahatma Gandhi Rd, Craig Park Layout, Ashok Nagar, Bengaluru, Karnataka 560001 | No Match: The suggestion matches on the location Modifier part of the POI. |
Prominence Match 4.1.2
Query String | Possible Suggestion | Explanation |
Ahmedab | Sardar Vallabhbhai International Airport, Ahmedabad | Match: A possible Intent for the query string is “Ahmedabad” The suggestion is highly prominent Transit POI located in Ahmedabad. This can be considered a match through prominence. |
CG Road, Ahmedabad, | No Match: A possible Intent for the query string is “Ahmedabad”. The suggestion is locally known shopping destination. However, it has no international prominence and would not be useful to the user potentially looking for the locality “Ahmedabad”. |
Abbreviation/ Alternate Name 4.1.3
Query String | Possible Suggestion | Explanation |
bom | Chhatrapati Shivaji Maharaj International Airport, Mumbai | Match: The international code for this airport is BOM. |
Category Match 4.1.4
Query String | Possible Suggestion | Explanation |
foo | Tomato’s Mardia Plaza, 1, 2, 3, Chimanlal Girdharlal Rd, Ellisbridge, Ahmedabad, Gujarat 380006 | Match: A suggestion for Multicuisine Restaurant is considered a match for the query foo |
metro station | Hazrat Nizamuddin Railway Staion, Nizamuddin East, New Delhi | No Match: The user is looking for a category metro station and the result is long distance railway station. Thus, not considered a match. |
General Match – Rating Example
Query locale: en_IN
Query: malad w
Viewport age: Fresh
Viewpoint centre: 18.98696248188791, 72.82841276755431
User location: 19.004167696153615, 72.81794142427263
1) Malad West Mumbai, Maharashtra, India Type: Address Lat, Long 19.192049270923693, 72.81030188253119 | Excellent- As Per GL 4.3.1 and 4.1.1 This suggestion matches the query string, and a Prominent Sub locality, therefore rated as EXCELLENT. |
1) Malad East Road Malad East, Mumbai, Maharashtra, India Type: Address Lat, Long 19.192049270923693, 72.81030188253119 | Bad- The user has entered a second token as ‘w’. It does not match with suggestion title. Since the second token is important, even though it is within the FVP rated as BAD. |
Distance/Prominence
Once you’ve decided how relevant the suggestion is based on user intent, consider the suggestion in the context of all possible suggestions in the real world. Each suggestion is given a rating based on prominence and distance to the user or viewport. The table below shows how they interact: Far distance and low prominence receive low ratings, while prominent suggestions close to the user or viewport receive high ratings.
Distance Prominence | Close | Medium* | Far** |
High | Excellent | Good | Acceptable |
Medium | Excellent to Good | Good to Acceptable | Acceptable to Bad |
Low | Excellent to Acceptable | Good to Bad | Bad |
*Medium distance away compared to similar prominent suggestion
**Far distance away compared to similar prominent suggestion
Distance/Prominence – Rating Examples
Query locale: en_IN
Query: Mumbai
Viewport age : Fresh
Viewpoint centre: 18.465450575507777, 73.78473741265023
User location: 18.483685822072307, 73.98386459636734
1) Mumbai Pune Expressway Lonavla, Maharashtra, India Type: Address Lat, Long 18.7693286859155, 73.40433582349463 Distance from VP: 52 Km Distance from User: 69.2 Km | As per GL 4.3.1 and 4.1.1 the suggestion title matches with query string. But as we can see the Suggestion is Moderate prominence and also very far from the FVP. As we refer to Distance/prominence table, the appropriate rating should be Acceptable. |
2) Mumbai Pune Bypass Road, Warje, Pune Type: Address Lat, Long 18.485831674278888, 73.79800612848496 Distance from VP: 2.63 Km Distance from User: 21 Km | As per GL 4.3.1 and 4.1.1 the suggestion title matches with query string. Although the result is low prominence, it is very close to FVP. Thus, appropriate rating for this suggestion will be – Excellent https://goo.gl/maps/tHCRCAqqM4vhkMvE7 |
Distance/Prominence – Rating Examples
Query locale: en_IN
Query: Ahmedabad
Viewport age: Stale
User location: 23.169031530748722, 72.58431354014459
1) Ahmedabad Gujarat, India Type: Address Lat, Long 23.019280235892296, 72.58170781130757 Distance from User: 16.61 Km | The suggestion is very well known and prominent locality. It should be considered Primary intent for the locality.Rating should be Excellent (Gl 4.3.3) |
2) Ahmedabad Railway Station Kalupur, Ahmedabad Type: Business Lat, Long 23.019280235892296, 72.58170781130757 Distance from User: 15.83 Km | The suggestion is not as primary as Locality Ahmedabad but is most prominent traint station for the city of Ahmedabad and Potentially satisfy the secondary user intent. There the relevance rating should be Good (User Intent) (GL- 4.1.2 and 4.2)
|
3) Ahmedabad University Commerce Six Road, Navarangpura, Ahmedabad. Type: Business Lat, Long 23.019280235892296, 72.58170781130757 Distance from User: 14.92 Km | As we are not sure if the query string is complete or incomplete. The user may add next token to the query string. Also, suggestion is moderately prominent and far from the user location. When we research in real world, we can see other prominent results which match the query string are present near user location. Thus, we can apply rating of Good/Acceptable (Distance/Prominence) for this result.(GL- 4.3.3) |
Locality suggestions and Prominence Match (GL- 8.1.4)
Query locale: en_IN
Query: Vas
Viewport age: Stale
User location: 23.169031530748722, 72.58431354014459
1) Vastral Ahmedabad Type: Address Lat, Long 23.019280235892296, 72.58170781130757 Distance from User: 5.35 Km | This suggestion is highly prominent Sub locality within Ahmedabad and closest to the user. This suggestion is most likely Intent for query string. Thus considered a prominence match. Therefore rated as Excellent. |
2) Vastrapur Ahmedabad Type: Address Lat, Long 23.019280235892296, 72.58170781130757 Distance from User: 12.33 Km | This suggestion is also higly prominent sub locality within the Ahmedabad city, but further away from the user location. This demoted to Good for distance/prominence issue. |
3) Vastral Road Gujarat, India Type: Address Lat, Long 23.019280235892296, 72.58170781130757 Distance from User: 6 Km | Although this suggestion is low prominence road but considerably closer to user, potential match. Thus, should be demoted to Good for distance/prominence. |
Distance Demotions – Rating Example
Query locale: en_IN
Query: Starbucks banas
Viewport age: Fresh
Viewpoint centre: 12.928651990889199, 77.52364245540304
User location: 12.923734076773709, 77.5504299977323
