From Excel Chaos to Real-Time Insights: The System Every Ecommerce Brand Needs
The ₹30 lakh per month ecommerce brand typically has more data than it can use and less insight than it needs. Revenue data in Shopify. Ad performance data in Meta Ads Manager and Google Ads. Inventory data in a WMS or warehouse spreadsheet. Marketplace revenue data in three different seller portals. Returns data in the OMS or in a separate spreadsheet someone maintains. Unit economics calculations in a master spreadsheet that two people have edit access to and that was last verified three weeks ago. When the founder wants to know whether the business is healthy whether this week's ROAS decline is a campaign problem or a product problem, whether the inventory position will support next month's planned marketing spend, whether the new SKU launched last month is performing they ask someone to pull the data, which takes half a day, by which point the situation may have already resolved or worsened. This is not a data problem. It is a data architecture problem. The data exists. The system to make it visible, current, and actionable does not.
If your business decisions are being made on data that is 3 days old, assembled from four spreadsheets, by someone who is not sure which version is current you are not running a data-driven business. You are running on intuition with the appearance of data. Here is the system that fixes this.
The Five Data Layers Every Ecommerce Brand Needs Connected
A real-time ecommerce intelligence system requires five data layers to be connected and updating continuously. The first is the commercial layer revenue by channel, by SKU, by customer segment, by geography, updated from Shopify/OMS/marketplace APIs on a sub-hourly basis. This is the layer most brands have in some form, though rarely at the granularity required to act on it. The second is the acquisition layer ad spend by platform, campaign, and ad set, with cost-per-click, conversion rate, and CAC by channel, updated from Meta Ads Manager and Google Ads APIs on a daily or faster basis. This layer is commonly available but rarely joined to the commercial layer in a way that allows true attribution analysis.The third is the inventory layer current stock by SKU by warehouse location, days-of-cover remaining based on current sell-through velocity, and incoming stock from purchase orders in transit. This layer is typically the most poorly connected it lives in a WMS or a warehouse spreadsheet that may update daily or may update whenever someone remembers to update it. The fourth is the financial layer cash position, accounts receivable from marketplace settlements, accounts payable to suppliers, and working capital headroom. This layer is typically the most siloed it lives in a Tally or Zoho Books instance that the accountant accesses once a week. The fifth is the fulfilment layer dispatch queue by day, NDR rates by courier and geography, average delivery time, and return initiation rates by SKU and channel.
Why Excel Fails as the Integration Layer
Excel and Google Sheets fail as the primary data integration layer for ecommerce brands above ₹20 lakh monthly revenue for three specific reasons. First, manual update dependency: any spreadsheet that requires a human to pull data from source systems and paste it into a master sheet is never more current than the most recent update, which is never more frequent than the human's available time. At 500 orders per day, the inventory spreadsheet that was accurate at 9am is meaningfully inaccurate by noon. Second, version control failure: shared spreadsheets with multiple editors generate the classic problem of multiple 'master' versions, conflicting numbers, and the unresolvable question of which version is current. Third, no alert infrastructure: spreadsheets display data but do not proactively alert the right person when a metric crosses a threshold that requires action. The CAC that crossed the profitability threshold yesterday is visible in the spreadsheet, but only if someone looks at the spreadsheet with enough frequency and attention to notice.
The Architecture That Works
The data architecture that solves these problems has three components. A data pipeline layer tools like Airbyte, Fivetran, or Stitch (₹5,000–₹20,000/month) that automatically extract data from every source system (Shopify, Meta, Google Ads, WMS, Tally) on a configured schedule (hourly for commercial and acquisition data, daily for financial data) and load it into a central data warehouse. A data warehouse layer Google BigQuery (pay-per-query, typically ₹2,000–₹8,000/month for ecommerce brands at this scale) or a hosted PostgreSQL instance that stores all the connected data in a single, queryable location with a consistent schema. A visualisation layer Looker Studio (free, Google's data visualisation tool) or Metabase (open source, free to self-host or ₹3,000/month hosted) that builds dashboards from the warehouse data and updates automatically as new data arrives.The total cost of this architecture for a ₹30–80 lakh monthly ecommerce brand: ₹10,000–₹30,000 per month for the full stack, or ₹5,000–₹15,000 per month if the data pipeline is replaced with Zapier/Make automations that write to BigQuery directly. The implementation cost: ₹50,000–₹1.5 lakh for a data engineer or analyst to build the initial pipelines and dashboards. The ROI: available in the first month from a single decision made on current data rather than stale data whether that is a campaign paused 24 hours earlier than it would have been, a reorder placed before a stockout, or a margin-destroying CAC trend spotted and addressed two weeks before it would have shown up in the monthly P&L.
The Five Dashboards Every Ecommerce Brand Should Have Live
| Dashboard | Key Metrics | Update Frequency | Primary User |
|---|---|---|---|
| Daily Commercial Health | Revenue vs target by channel, orders, AOV, conversion rate | Hourly | Founder, marketing lead |
| Acquisition Performance | CAC by channel, ROAS by campaign, spend vs budget | Daily | Founder, performance marketing manager |
| Inventory Position | Stock by SKU, days-of-cover, reorder alerts, dead stock flag | Daily (live for critical SKUs) | Operations lead, founder |
| Unit Economics | Net margin by SKU after returns and fulfilment, LTV by cohort | Weekly | Founder, finance |
| Fulfilment Health | NDR rate by courier and geography, return rate by SKU, CSAT | Daily | Operations lead, founder |
The Alert Layer: When Dashboards Are Not Enough
Dashboards solve the problem of data visibility for people who are actively looking. They do not solve the problem of notifying the right person when a metric crosses a threshold while they are in a meeting, travelling, or focused on something else. The alert layer configurable thresholds on key metrics that trigger WhatsApp or email notifications is what converts a data architecture from a reference tool into an operational intelligence system. The specific alerts that deliver the most value: CAC above the profitability threshold for any campaign with more than ₹5,000 daily spend; any SKU dropping below 14 days of cover; NDR rate in any geography crossing 25% on a 7-day trailing average; daily revenue below 70% of target at 6pm (allowing same-day intervention if a technical issue is suppressing conversion); and settlement variance above ₹5,000 from any marketplace. These alerts do not replace the dashboard. They ensure that critical information reaches the decision-maker in time to act on it, regardless of whether they are actively monitoring the dashboard.