O P T I M I S I N G S E A R C H A N O N -T E C H N I C A L P E R S P E C T I V E P A R T O F A n d r e a s B r u c k n e r S e n i o r M e r c h a n d i s i n g C o n s u l t a n t
Just some of the brands and retailers we are helping deliver relevant experiences to their shoppers Luxury brand 1 & Luxury Brand 2
1 . B U I L D A D E D I C AT E D S E A R C H T E A M
1. BUILD A DEDICATED SEARCH TEAM Merchandising Team Tech Team Good at: Know users and trends Know the shop features Should do: Own product performance Own infrastructure Own product data Own features 4
1. BUILD A DEDICATED SEARCH TEAM Merchandising Team Tech Team Good at: Know users and trends Know the shop features Should do: Own product performance Own infrastructure Own product data Own features If they own search: Make many small scale Make too few big changes changes and create and solve problems that problems that didn’t exist don’t exist 5
1. BUILD A DEDICATED SEARCH TEAM Merchandising Team Tech Team Good at: Know users and trends Analyse and generalise Know the shop features Should do: Own product performance Own user journey Own infrastructure Own product data (How to use shop features to find Own features the right products) If they own search: Make many small scale Make too few big changes changes and create and solve problems that problems that didn’t exist don’t exist 6
1. BUILD A DEDICATED SEARCH TEAM Ownership of search at top Fredhopper customers: No dedicated search expert 14% 7
1. BUILD A DEDICATED SEARCH TEAM Ownership of search at top Fredhopper customers: Dedicated search No dedicated search expert expert(s) 14% 19% 8
1. BUILD A DEDICATED SEARCH TEAM Ownership of search at top Fredhopper customers: Dedicated search No dedicated search expert expert(s) 14% 19% Dedicated finding expert 67% 9
2 . S E T R E A L I ST I C E X P E C TAT I O N S
2. SET REALISTIC EXPECTATIONS 100% 80% 60% Score 40% 20% 0% Precision Recall 11
2. SET REALISTIC EXPECTATIONS 100% Benchmark 80% 60% Score 40% 20% 0% Precision Recall 12
2. SET REALISTIC EXPECTATIONS 100% Benchmark 80% 60% Score 40% 20% 0% Precision Recall 13
2. SET REALISTIC EXPECTATIONS 100% Benchmark £1 m to £50m p.a. 80% 60% Score 40% 20% 0% Precision Recall 14
3 . B U I L D AWA R E N E S S A C R O S S T H E O R G A N I S AT I O N
3. BUILD AWARENESS ACROSS THE ORGANISATION Wrong data 16
3. BUILD AWARENESS ACROSS THE ORGANISATION Wrong data In Inconsistent data 17
3. BUILD AWARENESS ACROSS THE ORGANISATION Wrong data In Inconsistent data Mis isle leading data 18
4 . M E A S U R E T H E R I G H T T H I N G S
4. MEASURE THE RIGHT THINGS Search CVR 9.00% Started changes 8.00% 7.00% 6.00% 5.00% 4.00% 3.00% 2.00% 1.00% 0.00% 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 20
4. MEASURE THE RIGHT THINGS Signficant visits CVR Search CVR 9.00% Started changes 8.00% 7.00% 6.00% 5.00% 4.00% 3.00% 2.00% 1.00% 0.00% 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 21
4. MEASURE THE RIGHT THINGS Search NCVR Signficant visits CVR Search CVR 9.00% 160% Started changes 8.00% 140% 7.00% 120% 6.00% 100% 5.00% 80% 4.00% 60% 3.00% 40% 2.00% 20% 1.00% 0.00% 0% 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 22
4. MEASURE THE RIGHT THINGS Search PVR 1.00 Started 0.90 changes 0.80 PDP clicks per search session 0.70 0.60 0.50 0.40 0.30 0.20 0.10 0.00 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 Week 23
4. MEASURE THE RIGHT THINGS Search PVR 1.00 Started 0.90 changes 0.80 PDP clicks per search session 0.70 0.60 0.50 0.40 0.30 0.20 0.10 0.00 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 Week 24
4. MEASURE THE RIGHT THINGS • Online evaluation • Search result click rate • MRR of first click • Search refinement rate • Offline evaluation • Precision@1 • Precision@4 • Precision (all results) • Recall + additional low level metrics 25
5 . B U I L D A M E A S U R E M E N T F R A M E W O R K
5. BUILD A MEASUREMENT FRAMEWORK Exact match Spellcheck Partial match No results 5% 8% 13% 74% € 1.00 € 0.90 € 0.80 € 0.70 € 0.60 € 0.50 € 0.40 € 0.30 € 0.20 € 0.10 € - Value per search 28
6 . M I X O N L I N E A N D O F F L I N E E VA LUAT I O N
6. MIX ONLINE AND OFFLINE EVALUATION Weekly: • Find areas to Operational improve analysis Monthly: • Evaluate tactical performance analysis • Evaluate Ad-hoc impact of analysis new features Bi-annually: • Find pattern strategic • Find shifts in behaviour analysis 31
6. MIX ONLINE AND OFFLINE EVALUATION Weekly: Online evaluation • Find areas to Operational improve analysis Monthly: • Evaluate tactical performance analysis • Evaluate Ad-hoc impact of analysis new features Bi-annually: • Find pattern strategic • Find shifts in behaviour Offline evaluation analysis 32
7 . E C O M M E R C E I S D I F F E R E N T
7. ECOMMERCE IS DIFFERENT dress Name Name Summer dress Evening dress Category Category Casual dresses , Evening wear Beach dresses Style Style Formal Skater dress 34
8 . STA R T W I T H U S E R N E E D S
8. START WITH USER NEEDS 90% 80% 70% searches 60% of sea 50% ercentage of 40% 30% Per 20% 10% 0% Product Type Brand Product Features Individual Product 37
8. START WITH USER NEEDS 90% 78% 80% 70% searches 60% of sea 50% ercentage of 44% 40% 34% 30% Per 20% 14% 10% 5% 0.2% 0% Product Type Brand Product Features Individual Product Gender Size 38
9 . D O N ’ T F I X W H AT C A N ’ T B E F I X E D
9. DON’T FIX WHAT CAN’T BE FIXED 41
1 0 . Q UA N T I F Y A L L C H A N G E S
10. QUANTIFY ALL CHANGES Increase Add +£15m to revenue +16% to 28% +8% to 14% +3% to 6% searchable £30m annual UX: improve search PVR search CVR overall CVR attributes revenue search results +1% PVR = +0.5% 40% of revenue £500m annual CVR from search revenue 46
S U M M A RY
SUMMARY 1. Build a dedicated search team 2. Set realistic expectations (no 100% search quality) Organisation 3. Build awareness in the wider company (with buying team, etc.) 4. Measure the right things 5. Build a KPI framework Measurement 6. Mix online and offline evaluation 7. Ecommerce is different 8. Start with user needs, not features 9. Don’t fix what can’t be fixed. Optimisation 10.Quantify all changes 48
C O N TA C T U S F O R G E N E R A L E N Q U I R I E S sayhello@attraqt.com www.attraqt.com +44 (0) 20 3675 7800 C O N T A C T M E Andreas Bruckner abrueckner@attraqt.com
Recommend
More recommend
