Multi-Channel Reporting: Unifying Tokopedia, Shopee, and Lazada Data

Selling on multiple Indonesian marketplaces is normal. So is the spreadsheet someone updates every Monday morning to figure out which channel is actually working. The first is fine. The second is a sign you’ve outgrown what each platform’s native reports can give you.

Here’s what actually works to unify Tokopedia, Shopee, and Lazada data — without rebuilding it from CSVs every week.

Why platform-native reports fall short

Each platform has its own admin panel with its own reports. They’re fine for tool-specific questions (“how did my Tokopedia listing perform last week?”). They fail at the questions that matter most for cross-channel businesses:

• Which channel produces the highest contribution margin per product?
• Which products perform on one channel and bomb on another?
• Where are returns concentrated, and on which products?
• Which channel produces customers who reorder, and which ones don’t?

You can’t answer these by looking at three separate dashboards. The answers live in the joins between the data, not in any single platform’s view.

The shape of the right solution

Three layers, in order:

1. A small data layer you own

Pull data nightly from each platform’s API into a database you control. Postgres, Cloudflare D1, BigQuery, or any reasonable choice. The cost is small (~Rp 500rb–2 juta/month) and the value is enormous: you now have one place where every order, customer, and product lives, regardless of which platform it came from.

This is the part that’s almost always missing in companies that try to do multi-channel reporting in spreadsheets.

2. Identity matching across channels

The same customer appears in Tokopedia and Shopee with two different account names. The same product is listed slightly differently on each platform. Without matching these, the data layer is just three separate datasets sitting next to each other.

The pattern that works for products: a SKU master list maintained by your team, with each platform listing mapped to the SKU. Updated by hand when new products launch (one-time work per launch, not per platform).

For customers: matching by phone number plus first-letter-of-name is usually 80% accurate, which is good enough for analysis. The 20% mismatch is acceptable noise unless you’re doing precise customer-level work.

3. A dashboard on top of the data layer

Now you can answer the cross-channel questions. Metabase or Looker Studio, configured to query the data layer, gives you the cross-tabulations that the platforms’ native reports can’t.

What this typically produces: 6–12 charts that update automatically nightly. Channel performance per product, returns rate by channel, repeat customer rate by acquisition channel, contribution margin per channel after fees.

What gets surprising once you build this

Three patterns we see consistently for SMEs running multi-channel:

1. One channel is silently subsidising the others

You always knew Channel A has lower margins than Channel B. What you didn’t know: the gap is bigger than you thought, and Channel A is actually losing money once returns and fees are properly counted. Roughly 1 in 4 multi-channel SMEs we audit discovers this.

The fix is usually not abandoning the channel — it’s repricing or reformulating which products go on which channel. But you can’t do either until you can see it.

2. The same product performs very differently across channels

A product that’s a top seller on Shopee might be invisible on Tokopedia, and vice versa. The reasons (search algorithms, customer demographics, listing quality, price competition) take longer to figure out, but the visibility is the first step.

3. Returns are concentrated in specific patterns

Once you can see returns broken out by channel + SKU + customer, the patterns become obvious. Often: certain products on certain channels have 3x the return rate of the same product elsewhere. Sometimes it’s a bad product description; sometimes it’s a customer-base mismatch; sometimes it’s a fraud pattern.

What it costs to build

For an Indonesian SME selling Rp 1–10 miliar/year across 2–4 marketplaces:

• Data pipeline build (sync from each platform to a database, identity matching): Rp 40–100 juta, 4–8 weeks.
• Dashboard layer on top (Metabase configured, 8–15 key charts): Rp 15–40 juta, 2–4 weeks.
• Total upfront: Rp 55–140 juta, 6–12 weeks.
• Ongoing: Rp 1.5–4 juta/month (hosting, API costs, occasional maintenance).

Payback for SMEs at this scale: typically 3–6 months from time saved on manual reporting plus the operational decisions enabled by visibility.

Common mistakes to avoid

Three patterns that consistently fail:

Trying to skip the data layer

Tools that claim to “connect directly” from a dashboard to Tokopedia and Shopee work for the simplest cases and fall apart for cross-channel queries. The data layer is the part that lets you actually join across platforms. Don’t skip it.

Manual identity matching forever

The first version of identity matching is always partly manual. The fix to slowly automate it as patterns emerge. SMEs that never automate this end up with someone spending hours per week reconciling, which defeats the purpose.

Building once and never updating

Marketplace APIs change. New platforms (Blibli, TikTok Shop) become relevant. Your product line evolves. Plan for ~10% of the build cost annually as ongoing maintenance.

Where to start

If you’re already running on multiple marketplaces and the Monday-morning spreadsheet feels increasingly burdensome:

1. Audit your current reporting workflow. How many hours per week, how many people, what decisions are made off it. This is the value the new system replaces.
2. List the cross-channel questions you can’t currently answer easily. These define what the dashboard needs to do.
3. Decide whether to build it yourself or hire it out. Most SMEs hire it out — the data engineering work is specialised enough that internal teams struggle.

If you’re trying to figure out whether multi-channel reporting infrastructure is worth building for your specific situation, an hour of conversation usually clarifies it. We do those at no cost.