In this article, I walk you through my first research project on the search ads panel. We had to understand our users, who were professional digital marketers, to take the first step to improve our B2B product.

Introduction

Bazaar is the most famous Iranian app store, which provides different advertising channels for applications to reach more installs. Developers, managers, and digital marketers define ad campaigns in the Bazaars panel to hit their target. They specify keywords and set a price for each word. Whenever Bazaar’s end-users search for a word, the search ads service holds an auction, and the winner appears in the first stage of the search result. The amount determined for each installation that happens through search ads will be deducted from their accounts wallet.

Engineers had developed this service one year before, and it was working. Still, our organization invested more in this product and assigned several cross-functional teams to improve advertising products. I was a UX researcher in the advertising panel team.

We were at the beginning of the development and had numerous ideas. But before any action, it was necessary to understand our current users’ needs and problems.

Before planing

Before planning this research, I needed to familiarize myself with digital marketing terminology and learn how our product and our competitors work technically. So

• I took some digital marketing courses.
• I did a few interviews with digital marketers in our company.
• I reviewed previous documents about ads in the company and asked the team members about the terms I was unfamiliar with.

Goals and main questions

As a goal, we want users to use our product more and make unlimited campaigns. To reach the goal, we expected to answer these questions after researching:

• How do users and their companies think, plan, and decide about campaigns
• What makes them use or stop using an ad service?
• What are their severe pains and needs in the ad panel?

Method

Semi-structured interview: I chose this method because we needed deep qualitative insights about their context and organization and factors that affect the process of creating, optimizing, and evaluating campaigns. I would also ask them to show me their journey in ads services to make them recall the process. We did it with seven people from our target group.

Participants

We sought a segmentation to make our target user more straightforward. I had a conversation with our product manager and sales team, and in conclusion, We chose one segment based on sales segmentation as a starting point.

Interviewing

I had two categories of questions generally. one about their use of advertising services and the other about their context and organization. But according to users’ experiences, the interview took different paths, and I asked additional questions following the direction of the discussion. These are some examples of the interview questions:

About ads service

• Tell me about your last journey in the Bazaar ad panel; share your screen, and walk me through that.
• How did you create these campaigns? Share your screen and walk me through that.
• What were your worst and best experiences with Bazaar ads service? And why?
• How did you plan for these campaigns in your organization?
• How did you analyze and evaluate the results of that campaign?
• How did you report the result?
• If you have other experience with this service, tell me about them.
• If you use other channels for digital marketing, tell me about them too. How do you think those are different from Bazaar? Give me some examples.

About context and organization

• How did you choose between different channels in your organization?
• How did you plan your marketing?
• How many people influence decisions about campaigns, and how?
• How did you budget for campaigns?
• How do you compare different campaigns and channels?

Analyzing

I did seven interviews. Next, I wrote the transcript, Then started the thematic analysis.

1- I reviewed all data and generated initial codes. It was our first research, our question was broad, and we didn’t seek to analyze a specific theme, so I decided to do an exploratory analysis and coded based on the data.

2- I imported the relevant codes and raw data into Miro and placed the related data and codes close. Then define the themes.

3- I reviewed the theme and checked if there was a conflict in them or if it was too detailed or broad.

Prioritizing

We had about ten key insights and asked ourselves which were more important to address. So we used the following quick method to give us insight into users’ priority. I and the team’s product manager and designer answered these three questions about each user’s needs separately.

1. Does the problem occur on a red route (most essential user tasks)?
2. Is the problem difficult to overcome?
3. Is the problem persistent?

Then we shared our prioritization and discussed conflicts until we reached an agreement. Ultimately, we saw users’ needs in four prioritized groups: Low, Medium, Serious, and Critical.

Result and Findings

Here are some examples of prioritized user needs:

Here, I want to extract some of the most critical issues:

1- Campaign failure and not knowing why

Imagine users have planned to reach a specific installation at a particular time and were responsible for it. They choose Bazaar as a channel to achieve their target. But Sometimes users don’t get the expected result, and their budget was not entirely spent. It was stressful when they couldn’t hit their goal and didn’t know why or what to do. According to their experience, it was affected by several factors, like lack of insights about words and bids, lack of notification and guidance, and lack of facilities that increase the speed of installations.

2- Difficulty finding and choosing the word

We realized finding and setting words is one of the most challenging steps to creating and optimizing campaigns. Users didn’t know which words were searched most, which brought them more installation or revenue in Bazaar. They took time and tested each word to see if it was beneficial, but this testing process was too time-consuming and frustrating for them.

3- Not Getting Comprehensive Reports

They made several campaigns with different goals and targets. On the other hand, they needed to check some general metrics daily to ensure all campaigns were going well. Their managers were also required to know the overall status of the campaigns continuously every day.

We showed them detailed data for each campaign but did not give them a comprehensive view of their campaigns. They had to do calculations to see their daily metrics, which made evaluating campaigns too time-consuming, so they abandoned the measure. This problem sometimes causes them to notice late campaigns with low installs or other issues.

Solutions and Impacts

• We discussed the reasons for campaign failure and how confusing they are for users. Our goal wasn’t to help users understand why they failed. Instead, we wanted to reduce the possibility of campaign failure and increase the work speed. So our engineer started to work on smart ad service as a long-term solution. In this service, users determine the date of campaigns and budget; then, it chooses the best way to reach installation by artificial intelligence.

• We didn’t expect users to have so many challenges with choosing and setting words — which was crucial in search campaigns. We extracted all the factors that caused this problem from the interviews and started working on a keyword planner to quickly avoid the complexity of finding the right words. In this service, digital marketers define a few words, and our system shows them a list of suitable words. Those words were suggested based on the number of visits, conversion rate, and volume of competition.

• This research led us to several subsequent studies. In the future, we investigate which kinds of reports and metrics are helpful to show comprehensively. Also, we did some investigations to find out when users needed guidance and what type of guidance did they need? Then we redesigned the panels’ first view and also the notification system.
• The last and, in my opinion, one of the most critical impacts of this project was that all team members practiced seeing the product from the user’s point of view because they were involved in different project stages.