Gross Insurance

First, sharpen the axe

Problem

The site was slow and dated. The team was three back-end developers and one front-end developer, with no designer.

What I did

Proposed hiring a designer and three full-stack developers, found the candidates, interviewed them and chose whom to hire. Studied the site and its analytics to find the bottleneck. Made the site faster first, and in parallel designed the new version.

Result

By June there was a team that could ship, a faster site and a design ready to build. That is why the OSAGO redesign took four weeks.

Companies could not pay

Problem

The site took payment only through Payme and Click. Companies have to pay with a corporate card, so a corporate client could not buy a policy online at all.

What I did

Found a payment provider that takes corporate cards, Atmos, set up the meeting and negotiated, got management's approval. Card payment went live on the OSAGO page on 21 July.

Result

Companies can buy OSAGO on the site. A group of customers that could not buy online at all now can.

OSAGO sales were falling. The redesign turned them

Problem

OSAGO sales through the site had fallen six weeks in a row, and the last week before the release was the lowest of the period. Most visitors never got as far as the OSAGO page, and most of those who did left without buying.

What I did

Noticed the decline and decided what to change, drew the prototype for the designer and worked out the design together, set the priority and the scope. In four weeks, 2 to 29 June, the team rebuilt the OSAGO calculator and the whole path to it: the first screen, the products section, the navigation, the mobile versions: three visitors in four come from a phone. Three front-end developers, two back-end developers and a designer; refinements until 6 July.

Result

OSAGO policies a week through the site up 64%, and the level has held for seven weeks. The site's conversion to a purchase is the highest on record.

Counted as weekly OSAGO policies through the site, six weeks before the release against seven after; no advertising ran in those weeks. Where the growth came from: traffic grew 8.8%, the rest came from the funnel. A third of the gain came from the path to the calculator, not from the calculator itself. And more visitors are still lost before the OSAGO page than on it: that is the next lever.

Measuring the funnel step by step

Problem

The analytics counted visits and purchases, but not the steps between them, so nobody could say where the June growth came from.

What I did

Rebuilt the Yandex Metrica goals in mid-July so the funnel reads step by step: a visit, the OSAGO page, the order, the payment.

Result

The June result splits into traffic, the path to the OSAGO page and the page itself, and the next lever is visible: more visitors are lost before the page than on it.

Sign-in without a password

Problem

Customers signed in to the cabinet with a phone number and a password, and passwords got forgotten.

What I did

Moved sign-in to MyID, the national digital identity: live on 5 August, and from 10 August one MyID opens several accounts. Registration with a password went away.

Result

Nothing to remember and nothing to reset: the cabinet opens with the same ID people use for government services.

The OSAGO calculator, a second pass

Problem

The June calculator was built in four weeks. On phones and at the details step it still needed work, and the price could show before the car was known.

What I did

Reworked the calculator for phones, kept the price hidden until the plate number is in, added an order summary beside the form and a new details step, and a new screen after payment. Live on 21 August; a new drivers step and a new travel calculator followed in early September.

Result

The price a customer sees is always the price for their own car, and the calculator was rebuilt for the phone screen.

The cabinet fills in the calculator

Problem

A customer with a cabinet typed in passports and car details again on every purchase. Booking a doctor under a health policy meant a phone call.

What I did

I made the cabinet the priority. A release on 11 September: the calculators suggest the people and cars saved in the cabinet, a family section, and booking a doctor under the health policy online.

Result

Data saved once fills the next purchase, and one more office task moved online.

Claims and cancellation move online

Problem

To report an insured event or to cancel a policy, a customer had to come to the office.

What I did

Two cabinet releases on 30 September: report an insured event, follow it, see what is left of the health policy limits, cancel a policy online with the refund paid to the card.

Result

Gross was the first insurer in Uzbekistan to let customers report an insured event and cancel a policy online, with the refund paid to the card. The cabinet covers the policy from purchase to cancellation. These functions have not been announced yet, and customers already find them and file requests online every week.

The story on the blog: first in Uzbekistan, insurance without an office visit

A tracker the team actually opens

Problem

The team had Jira, but it never took: the developers did not open it, so there was no single order of work.

What I did

Built our own tracker, northly: every request in one place, the large ones scored by reach, impact, confidence and effort, the work grouped into releases with dates.

Result

The team uses it and likes it. Management has access too, so they see what we are working on without asking.

KASKO: the price without a call

Problem

To learn what KASKO cost, a customer had to call a manager. The KASKO page was a picture of three cars and one button.

What I did

My initiative: a calculator for KASKO. Type the car's value and see Silver, Gold and Platinum with a yearly price. A second pass on the OSAGO order form went out in the same release.

Result

The KASKO price needs no call. KASKO is rarely bought online, so the effect will be measured as the quarter after the release against the quarter before.

A campaign in three days

Problem

People buy OSAGO because the law requires it, then leave. Nothing on the site led them from the mandatory policy to a voluntary one.

What I did

I had proposed a cross-sell offer like this many times; for Gross's 15th anniversary management agreed, and on 19 October we fixed the terms: buy OSAGO, get 15% off any voluntary policy until 31 December. The discount in the payment service, the offer on the home page and the OSAGO page, and counting of policies that come from the ads.

Result

Started on 22 October, three days after the decision, and runs until 31 December. The result will be counted when it ends.

Insurance inside Telegram

Problem

People in Uzbekistan live in Telegram, and buying a policy meant leaving it for the site.

What I did

My initiative: the Gross Mini App. OSAGO and travel insurance are bought without leaving Telegram, with the same calculators as the site.

Result

A new sales channel: policies are already sold through it.

The mobile app, audited and redesigned

Problem

The mobile app lagged behind the website: the cabinet on the site could do more than the app, and nobody had a full list of what was broken.

What I did

Catalogued all 78 screens with every bug found, then designed the new version myself, 24 screens with sign-in without a password by SMS code or MyID, and tested it.

Result

In development: the app is getting everything the cabinet on the site does.

Disclosure and security

Problem

As a joint-stock company, Gross must publish set information on its own website. The reports went to the official registry, not to the site.

What I did

On my initiative: checked the site against the 22 items the rules require, designed a disclosure section, and started a security audit of the site.

Result

The security audit was completed in November; the disclosure section is designed and waits for its data.