Customer Support

How to run nps and csat surveys as CPG Founders

Customer SupportFor CPG Founders3 apps12 steps~24 min to set up

You're getting customer feedback the worst possible way — a flood of one-star Amazon reviews you didn't see coming, a Shopify inbox full of 'where's my order?' that you answer manually at 11pm, and the occasional distributor complaint that surfaces six months after the product hit shelves. You don't run formal NPS or CSAT surveys because the tools built for that (Delighted, Medallia, Qualtrics) assume you have a CX team to analyze the results and act on them. So you're flying blind on what your retail buyers actually think, what's driving repeat purchase or churn on DTC, and whether your latest reformulation landed. That silence is expensive when your brand lives or dies on reviews and reorders.

Build this on Starch → See the apps used

Customer SupportFor CPG Founders3 apps12 steps~24 min to set up

Outcome

What you'll set up

An automated NPS survey that goes out to DTC customers 14 days after delivery and pipes responses into a dashboard you can actually read without a data analyst

A CSAT flow triggered after distributor or wholesale inquiries close — so you know whether your B2B relationships are healthy before a buyer decides not to reorder

A weekly digest that surfaces low scores, flags recurring complaint themes by product SKU, and drafts outreach to detractors — all without you manually sorting a CSV

The Starch recipe

Apps, data, and prompts

The combination of Starch apps, the data sources they pull from, and the prompts you use to drive them.

Apps used

CRM Customer Support Agent

Data sources & config

Starch syncs your Gmail data on a schedule to send surveys and log replies. Connect Shopify from Starch's integration catalog; the agent queries it live to detect delivered orders and pull order and SKU data. Connect Slack from Starch's integration catalog; the agent queries it live to send your weekly digest and urgent alerts. Survey responses and contact records are stored in Starch's CRM surface so every score is tied to a real customer and a real SKU.

Prompts to copy

Build me an NPS survey automation that triggers 14 days after a Shopify order is marked delivered. Send the survey via Gmail, log every response with the customer's name, order ID, product SKU, and score, and flag any score below 7 for follow-up. Show me a dashboard of scores by SKU and by month.

Every time an email thread in my Gmail tagged 'wholesale-inquiry' is marked resolved, send a 3-question CSAT survey to that contact. Log responses against the contact record in my CRM and surface any score below 4 to me immediately in a Slack message.

Every Monday morning, pull all NPS and CSAT responses from the past 7 days. Group complaints by theme — shipping, taste, packaging, price — and draft a reply for any detractor I haven't responded to yet. Send me a summary in Slack.

Run these in Starch → or paste them into your favorite agent

Walkthrough

Step-by-step

1 Connect Shopify from Starch's integration catalog. The agent will query it live to detect when an order status changes to 'delivered' — this is the trigger for your post-purchase NPS flow.

2 Connect Gmail as a scheduled-sync provider. Starch syncs your Gmail data on a schedule so it can send survey emails from your actual address and read responses as they come back.

3 Tell Starch: 'Build me an NPS survey automation that fires 14 days after a Shopify order is delivered. Email the customer from my Gmail account, include a 0–10 scale and one open-text field asking what drove their score.' Starch builds the flow.

4 Set up response logging: 'Log every NPS response to a CRM record that includes the customer name, email, Shopify order ID, product SKU, score, and verbatim comment.' Now every response is tied to a traceable order.

5 Set the detractor alert: 'Any NPS score below 7 should trigger a Slack message to me immediately with the customer name, SKU they bought, their score, and their comment.' Connect Slack from Starch's integration catalog so the agent can query it live to send the alert.

6 Build the wholesale CSAT flow: 'Whenever a Gmail thread tagged wholesale-inquiry is resolved, send a 3-question CSAT survey to the contact within 2 hours. Log responses to their CRM record.' This keeps your B2B relationship health visible.

7 Install the Email Triage app from the Starch App Store and tell it: 'Flag any inbound email that contains the words NPS, review, or complaint as high priority.' This makes sure survey replies and organic feedback don't get buried.

8 Build the SKU-level dashboard: 'Show me a dashboard of average NPS score by product SKU and by month, plus a breakdown of open-text complaint themes — shipping, taste, packaging, price — using responses from the past 90 days.'

9 Set up the weekly digest automation: 'Every Monday at 8am, pull all NPS and CSAT responses from the past 7 days. Group complaint themes, list every detractor I haven't replied to, draft a reply for each, and send the full summary to my Slack.' Review and send the drafted replies with one click.

10 For Amazon review monitoring — no official API needed. Tell Starch: 'Every week, check my Amazon seller page for new 1-star and 2-star reviews, pull the review text and ASIN, and add them to my NPS dashboard as a separate source.' Starch automates this through your browser — no Amazon API required.

11 After 60 days, ask Starch: 'Which SKUs have the lowest average NPS? Which complaint themes appear most often for each? Show me whether scores improved after my April packaging change.' Use the answer to prioritize your next reformulation or packaging decision.

12 When Customer Support Agent launches (currently in development — request beta access), wire it in to handle inbound responses automatically: it will resolve common follow-ups, escalate complex complaints to you with full thread context, and close the loop without you touching the inbox.

See this running on Starch

Connect your tools, describe what you want, and the agent builds it. Closed beta is free.

Try it on Starch →

Worked example

Q1 2026 NPS Audit — Oat Protein Bar Brand, 3 SKUs

Sample numbers from a real run

NPS responses collected (DTC, 14-day post-delivery)	214
Average NPS — Original flavor	61
Average NPS — Dark Chocolate flavor	43
Average NPS — Peanut Butter flavor (reformulated Feb 2026)	29
Detractor follow-ups drafted by Starch	38
Detractor replies sent within 48 hours (vs. prior quarter: 4)	31
CSAT responses from wholesale contacts (Sprouts, KeHE inquiries)	27
Average CSAT — wholesale B2B	4.1

In January 2026, the founder set up a 14-day post-delivery NPS flow through Gmail for all three SKUs sold on their Shopify DTC store. By end of Q1, 214 responses came in. The Original flavor held a 61 NPS — healthy. But the Peanut Butter SKU, which had been reformulated in February to cut costs on protein isolate, scored 29. Starch's weekly digest surfaced this pattern in week three of February: 14 out of 19 Peanut Butter detractors mentioned 'texture' or 'chalky' in their open-text responses. The founder saw this in the Monday Slack digest, used Starch's drafted reply to personally reach out to 11 detractors with a replacement offer, and brought the complaint data directly into a call with the co-packer to revisit the formulation. On the wholesale side, 27 CSAT responses from Sprouts and KeHE buyer contacts averaged 4.1 out of 5 — good, but two contacts scored 2, both citing slow response time on a mislabeled pallet dispute. Starch flagged those immediately via Slack at the time of survey completion. Without this system, the Peanut Butter reformulation problem would have shown up as an Amazon review problem three months later, after retail buyers had already seen the rating drop.

Measurement

How you'll know it's working

NPS score by SKU (tracked monthly, compared pre/post reformulation or packaging change)

Detractor response rate and time-to-reply (how fast you close the loop on unhappy customers)

CSAT score from wholesale and distributor contacts (leading indicator of reorder health)

Complaint theme frequency by SKU (shipping damage, taste, packaging integrity, price)

Review velocity on Amazon — 1-star and 2-star count per ASIN per week

Comparison

What this replaces

The other ways teams handle this today, and how the Starch version compares.

Delighted or Typeform + Zapier

Good standalone survey tools but they drop results into a separate dashboard that isn't connected to your CRM, SKU data, or Shopify orders — you still have to manually cross-reference scores with products and chase down follow-ups.

Gorgias

Built for support ticket volume, not structured NPS/CSAT measurement — you'd need to bolt on a separate survey tool and still do manual analysis to get SKU-level insight.

Qualtrics or Medallia

Enterprise-grade feedback platforms with real analytical depth, but they're priced for brands with a dedicated CX team and start around $15,000/year — not built for a 3-person CPG operation.

Manual Gmail + Sheets

Free and familiar, but you're writing every survey email yourself, copying responses into a spreadsheet, and doing your own theme analysis — which means it only happens when you have three hours and a quiet weekend.

On Starch RECOMMENDED

One platform — crm, customer support agent, founder inbox all running on connected data. Setup in plain English; numbers stay current via scheduled syncs and live agent queries.

Try it on Starch →

FAQ

Frequently asked questions

Can Starch actually send survey emails from my own Gmail address, not some generic noreply@?

Yes. Starch syncs your Gmail data on a schedule and sends outbound emails through your connected Gmail account. One thing to know: Gmail's OAuth consent screen currently shows the name of Starch's underlying email client rather than 'Starch' — this is on the roadmap to fix. For most CPG founders sending to existing customers, this hasn't been a deliverability issue.

What if my customers buy through Amazon, not Shopify? Can I still collect NPS?

Amazon doesn't allow sellers to contact buyers directly for NPS, so you can't trigger a post-purchase email from order data there — that's an Amazon policy limit, not a Starch limit. What Starch can do is automate monitoring of your Amazon reviews through browser automation (no Amazon API needed) and pull new 1- and 2-star reviews into your feedback dashboard weekly so you're at least tracking sentiment by ASIN.

Will this work if I sell through distributors like KeHE or UNFI and don't have direct customer emails?

For end consumers, no — if the distributor owns the retail relationship, you don't have buyer emails. But you do have your buyer contacts at the distributor and at retail chains. The CSAT flow works well for those B2B relationships: wire it to trigger after any email thread with a buyer contact closes, and you get a structured read on whether your wholesale relationships are healthy before someone decides not to reorder.

Can Starch analyze open-text NPS comments and group them by theme automatically?

Yes. Tell Starch what themes matter to you — packaging, taste, shipping, price, ingredient concerns — and it will categorize open-text responses against those themes in your weekly digest. You can also ask it ad hoc questions like 'what are the most common complaints about my Peanut Butter SKU in the last 60 days?' and get a real answer from your actual response data.

Is there a pre-built NPS app in the Starch App Store I can start from?

There's no dedicated NPS starter template today — but the CRM app and Email Triage app (both live in the App Store) are the right starting surfaces. You describe your survey workflow to Starch and it builds the automation on top of those connections. Most CPG founders are up and running with a working NPS flow in under an hour.

What about the Customer Support Agent I saw mentioned — can that handle survey follow-ups?

Customer Support Agent is currently in development. When it launches, it will be able to handle inbound replies to your NPS surveys automatically — answering common follow-up questions, escalating detractors with full context, and closing the loop without you in the thread. You can request beta access today to get notified when it's available. For now, Starch drafts the follow-up replies for you and you send them with one click.