erpkaizen

Measuring the value

Trying Zalo-in-ERP with no sign-up: in the real inbox after 2 seconds — and nobody stayed 30

Has a measurement By 9 Nov 2026 About 9 minutes
Left: a phone showing the ERP inbox, a sample Zalo message just arriving at the end of the chat. Right: two time bars — a demo takes days, the self-serve trial takes seconds. No real data. Inbox Quote for… Just arrived Sales demo days Self-serve trial seconds until the first look
A demo: days until the first look. A self-serve trial: seconds, inside the real inbox.

A trial page puts the visitor straight into the real ERP inbox, a sample message just arriving, no form, no password. 43 visits from ads: 35 reached the inbox within 2 seconds, 5 stayed 10 seconds, none stayed 30. What a demo costs, and how much of it a trial can replace.

What does a sales demo cost, and how much of it can a self-serve trial replace?

Enter the people asking for a demo each month, the salesperson's minutes per demo, the cost of an hour, and the share of people for whom the trial is enough to decide on their own. The trial lane on the chart is measured; the demo lane is yours.

People who asked — not strangers from an ad
Scheduling, preparation, travel or call, the demo itself
Salary + insurance + average commission, per hour
Deciding to buy and deciding not to both count. Illustrative — not yet measured

Derived

One demo that actually takes place (VND)–
One visit to the trial (VND)–
Demos no longer needed per month–
Wait until the first look–
Assumptions — click to see and change each number

The demo route

The trial route

How long until the first look, and how many are still there Logarithmic time axis: each tick is 10× or more the one before · dot area is the share still there · a hollow dot is nobody · hover for the numbers
How long until the first look, and how many are still there

A. Salesperson hours not spent on demos–
B. Cost of bringing people to the trial–
Per year, in money–

This is an estimate, not a promise. The field that decides everything is the share who decide after trying — illustrative, not yet measured. Set it to 0 and the cost per visit to 0: the total goes to 0. An honest model has to be able to do that.

Business software is usually sold through a demo: an interested person leaves a phone number, someone calls back, a date is set, and a few days later there is an hour of watching a screen. A demo costs the salesperson hours, and it costs the buyer days of waiting — before they can see whether the product is what they need at all.

This piece measures a different way for one small feature — a Zalo inbox inside the ERP, next to customer and supplier records: one click and the visitor is standing in the real inbox, no sign-up, no password. The question: how fast does the first look arrive, how many people stay, and how much of a demo can a self-serve trial replace? Real numbers, including when they are small.

1. What changes

The trial page is built so that value comes before any action:

  1. Opening the page starts a trial session with a code — no form, no email, no password.
  2. Less than a second later the page moves into the ERP's real inbox screen under a temporary account belonging to that session alone. There is no simulated screen in between.
  3. As the visitor arrives, a sample Zalo message "just arriving" sits at the end of the chat, unread. Its content matches the ad that brought them: someone who came from the ad about supplier promises sees a supplier asking to push a delivery back; everyone else sees a customer asking for a quote.
  4. A five-step on-screen guide points at the right places: the message that just arrived · the button that opens the sender's record and orders · the reply box · the list of all chats · how to try with your own Zalo (send the trial code to a Zalo account; your message appears in the inbox a few seconds later).
  5. The sample chat is labelled: replying here does not send anything to the real Zalo. Advanced features are shown but locked.

The page measures time spent in seconds while the screen is showing — a mark at 10, 30 and 60 seconds — and every step the visitor takes.

2. Measured: 43 visits from ads

From the moment the new page went live (01:45 on 30/10/2026) until 06:50 on 31/10, every visit came from ads (counted in the page's own trial-session log, totals only):

  • 43 visits. 42 of them on a phone, inside the Facebook app's built-in browser.
  • 35 were let into the real inbox within 0–2 seconds of opening, by the server's clock — the sample message was written into the chat at the same moment. 8 did not get that far.
  • 5 stayed 10 seconds with the screen showing. Their 10-second mark reached the server 12–24 seconds after they were let into the inbox, meaning the inbox screen needed roughly another 2–14 seconds to appear on their phones.
  • None stayed 30 seconds. Nobody replied to the sample message, attached the chat to a record, tried with their own Zalo, asked for advice or signed up.

The ads in those hours cost about 38,000 VND, roughly 900 VND a visit (the ad platform's figures, read on the morning of 31/10). The previous page — where a visitor had to copy a code, open Zalo and message an unknown number before getting in — had 45 visits from ads; none clicked through to Zalo, none reached the inbox. The two pages measured different things, so only one comparison holds: the old page brought nobody to a first look.

Read plainly: the new design solved "how long until the first look" — seconds instead of days. It has not solved "does anyone stay". Most visitors were people scrolling Facebook who tapped an ad, not people looking for software. For them, an open inbox is not yet reason enough to stay 30 seconds.

3. The first stream: salesperson hours not spent on demos

A self-serve trial does not replace a demo for strangers; it can replace one for people who already asked. Send the trial link to people who ask for a demo before booking it, and some of them will decide on their own — to buy or not — without the demo.

Money/year = demo requests per month × share who decide after trying × minutes per demo ÷ 60 × cost of an hour × 12
Example 30 demo requests a month. A demo costs the salesperson 90 minutes including scheduling, preparation and travel; 120,000 VND an hour — 180,000 VND a demo, or 225,000 VND per demo that actually takes place if 20% of bookings do not show. Suppose 25% decide after trying. 30 × 25% × 90 ÷ 60 ≈ 11 hours a month, 135 hours a year ≈ 16.2 million VND.

How to measure it. For one month, send the trial link to everyone who asks for a demo, before calling back. Count how many still book the demo and how many decide without it. Time 10 real demos, preparation included.

How far to trust it: low, because the deciding figure — 25% — is illustrative. We have not measured it: the 43 visits above were strangers from ads, not people who had asked.

4. The second stream, with a minus sign: the cost of bringing people to the trial

Money/year = demo requests per month × trial visits per request × cost per visit × 12
Example Each person asking for a demo opens the trial once, at 900 VND a visit — the figure measured when visits are bought with ads. 30 × 1 × 900 × 12 ≈ 324,000 VND a year.

If the link is sent by hand to people who asked, this is close to zero. It only grows when visitors are bought with ads — and then the thing to divide by is not visits but people who stayed: 38,000 VND for 5 people staying 10 seconds is about 7,600 VND each, and for people staying 30 seconds it cannot be divided yet.

5. Adding up

StreamPer year
Salesperson hours not spent on demos (135 hours)16.2 M
Cost of bringing people to the trial−0.3 M
Total15.9 M

About 16 million VND a year — if 25% decide after trying. Set that figure to 0, which is what the measurements so far say, and the calculator is left with a small negative: the trial is only a cost. Those 135 hours are money or capacity for the same salesperson to meet more people, never both.

One thing the calculator does not turn into money: the time to the first look. The demo lane in the chart above takes about three days; the trial lane takes two seconds. The buyer does not have to wait to find out whether this is for them.

6. What this ROI rests on

The share who decide after trying. Everything stands on it. A trial people open and close within 5 seconds replaces no demo, however cheap it is.

Who is opening it. Strangers from ads and people who already asked are two different groups. This piece measured the first; the ROI in section 3 needs the second. Do not use one group's rates to guess the other's.

The screen has to appear in time. On a phone the inbox screen needs a few more seconds after the server has opened the door — during which the visitor sees a page loading. Shortening that gap is the first thing worth measuring again, because it comes before every other step.

7. Measuring it in your own company

  1. Baseline the demo. For one month: demo requests, demos that took place, minutes per demo including preparation, days from the request to the demo.
  2. Measure the trial in seconds, not visits. Mark 10, 30 and 60 seconds while the screen is showing, and every action. A visit on its own is nothing yet.
  3. Test it on people who asked. Send the link before calling back, then count as in section 3. Only that number says how much of a demo the trial replaces.

8. Worth remembering

Dropping the sign-up form and putting the visitor straight into the real product cut the time to the first look from days to seconds — at less than a two-hundredth of the cost of a demo. It does not make anyone stay by itself. The trial measured that on its first day, and a measured zero is more useful than a guessed rate.

The time to the first look is something design can solve. The reason to stay has to be measured on real people.

Visits, visits reaching the inbox and their timing, visits staying 10 and 30 seconds, actions, device types and ad cost were measured on ERP Kaizen's own trial page from 01:45 on 30/10 to 06:50 on 31/10/2026, totals only. Demo requests, minutes, wages, the no-show rate and the share who decide on their own are illustrative. Replace them with your own.

Read next