The success of your personalisation effort is wholly dependent on data? We recommend you start-small & use accurate data that already exists.
To see read from the start:
🔗 Personalisation Value Chain
Data (2 of 6)
1️⃣ Data > The New Oil
→ Hype around all these different AI models.
→ All Models are 100% dependant on accurate data.
→ Interaction Data needs to include relevant identifiers.
2️⃣ Define Ideal Personalisation Profile
→ When doing your customer journey mapping – identify required attributes (demographics, product holdings, purchase history, loyalty data, interactions, self-care usage, intent, etc).
→ finalise ideal profile requirements (incl. the identifier used to request profile).
3️⃣ Shifting from Data Silos to a Centralised Data Lake
→ Big data programs take years to complete (& often fail).
→ Don’t tie the success of your personalisation program to these initiatives (they may also have different priorities).
→ Formally hand-over your ideal profile requirements (desired attributes & draft api specifications) & move on.
4️⃣ Start small & use an existing data attribute
→ Initially, only focus on 1 profile element (e.g. location, traffic source or browser/device type).
→ Start small – so that only a few creative variations are required.
→ Identify issues before a significant investment is committed.
→ You may also uncover existing (free) capability that you didn’t know existed (more on this in future posts).
5️⃣ CRM/CDP Data
→ You will never use personal (PII) data for personalisation.
→ First-name will be used > this is not considered as PII unless it can associated to other profile data (e.g. last name, address, IP address).
Decision needs to be made about what the source of truth will be for your personalisation profile (remember this needs to support both customers & prospects & must support real-time).
—————
In my next post, I’ll cover 🔗 Personalisation Value Chain: Models (3 of 6)
————–
