targeted marketing
play

Targeted Marketing By Geomatics Development & Services, Telekom - PowerPoint PPT Presentation

Location Analytics for Targeted Marketing By Geomatics Development & Services, Telekom Malaysia Berhad Introduction Location Based Information critical in TM. Sales & Marketing Upsell activities Sales forecast Operation


  1. Location Analytics for Targeted Marketing By Geomatics Development & Services, Telekom Malaysia Berhad

  2. Introduction Location Based Information critical in TM. • Sales & Marketing • Upsell activities • Sales forecast • Operation • Reduce waiting time • Avoid cable cut • Planning • TM Point outlet • Webe tower planning

  3. TM G eomatics Introduction • Established since 1992 • Responsible for TM GIS Map for TM internal and external use • Manpower: 80 persons

  4. TM Marketing Campaign

  5. Sales Target / Forecas t Property No of Daerah density exchanges Ampang 1594 4 Kuala Lumpur 773 16 Petaling 593 20 Sabak Klang 362 13 bernam Ulu Langat 191 9 Ulu Selangor Gombak 179 12 Putrajaya 116 1 Kuala Selangor Sepang 65 5 Gombak Kuala Langat 56 9 Hulu Selangor 41 12 KL Kuala Selangor 37 5 Ulu Petaling Sabak Bernam 18 4 Langat Klang P U J Sepang Kuala Langat

  6. Fas ter Service Order • Ease the front liner to identify infra availability & capability • To response customer immediately on infra status

  7. Fas ter Service Order BEFORE AFTER 66% 34% 96% 4% 68 68 45 65 23 3 Total order Kajisiasat Total order Waiters order success order success

  8. Fas ter Service Order

  9. Fas ter Service Order

  10. Avoid Cable Cut

  11. TM Point & Authorized D ealers Coverage Planning

  12. TM Point & Authorized D ealers Coverage Planning

  13. Webe Tower Planning

  14. Webe Tower Planning X: numbers of Unifi customers Y: numbers of Streamyx customers Z: distance from the tower

  15. Webe Tower Planning Low density area High density area Eg: Kuala Selangor Eg: Ampang Density 37 property per sq km Density 1594 property per sq km Distance from one exchange to another ~ 8 to 10 km Distance from one exchange to another ~ 3 to 4 km

  16. Challenges D oing Location Bas ed A nalytics • Customers’ addresses are not clean and standardized • Accuracy of Geocoding • Time to process large amount of data

  17. Cleaning & Standardizing A ddres s es 17EOIZPHASE 2KOTA WISMA TUNE NO19 KINABALU INDUSTRIAL LORONG DUNGUN PARK JALAN DAMANSARA HEIGHTS SEPANGARMEN KOTA 68100 KUALA LUMPUR KINABALUSAB merge street No comma 11, JALAN BAKAWALI 69 , 11, JALAN BAKAWALI 69, TAMAN JOHOR JAYA , 81100 TAMAN JOHOR JAYA, 81100 JOHOR BAHRU JOHOR BAHARU With comma Different spelling

  18. Cleaning & Standardizing A ddres s es 11, TAMAN JOHOR JAYA, 11, JALAN BAKAWALI 69, 81100 JOHOR BAHRU 81100 JOHOR BAHARU Missing street name Missing section name PEJ PENGARAH TANAH LOT PTD 119913 (NO. 23 DAN GALIAN JOHOR, ARAS JALAN NB 2 1/1), TAMAN 5 BGN SULTAN IBRAHIM, NUSA BISTARI 2, 81300, JLN BKT TIMBALAN, 80000, SKUDAI, JOHOR, JOHOR BAHRU, JOHOR Lot tanah acronym

  19. Cleaning & Standardizing A ddres s es Dirty & Non Clean up & standardized standardize address Cleaned & standardized Address Dictionary Example: Jln-> Jalan Lrg -> Lorong Lebuhraya Mahameru -> Lebuhraya Sultan Iskandar

  20. Cleaning & Standardizing A ddres s es Clean data by group to reduce manual intervention

  21. G eocoding~ Location A ccuracy 2 1 Property level 1 2 Street level 3 Section level 3

  22. G eocoding ~ A ccurately G eocoded? • TM needs high accuracy in geocoding / text matching • If it is wrongly geocoded, the analysis results will be wrong as well Accuracy: (100 - Levenshtein distance) __________________________ x 100% Number of letters in a word

  23. G eocoding ~ A ccurately G eocoded? • Levenshtein Distance • Between two words is the minimum number of single-character edits to change one word into the other. • Jalan <> Jln = 2 • Enterprise <> Entreprise = 2 • MA Sdn Bhd <> ML Sdn Bhd = 1

  24. G eocoding ~ A ccurately G eocoded? • TM only accepts if accuracy > 88% . • Manage to geocode 70% - 80% of addresses. • With less manual intervention .

  25. Time to Proces s H uge A mount of D ata 1 Worker process Data to be Processed processed data Data size: 300k Time required: 24 hours

  26. Time to Proces s H uge A mount of D ata 10 worker processes Data to be Processed processed data Data size: 300k Time required: 4 hours

  27. Time to Proces s H uge A mount of D ata Scale up the system easily Data to be Processed processed data

  28. Benefits • Increase Revenue • accurate planning analysis • better customers coverage • Increase Customers Satisfaction • less waiting time • Save Costs • cable cut

  29. Thank You

Recommend


More recommend