
In the last couple of lessons, we identified our target audience, recruited them, and conducted interviews.
在過去的幾節課中,我們確定了目標受眾,招募了他們,並進行了訪談。
In this lesson, we’ll discuss how to debrief and synthesize user interviews into clear, valuable insights. This is a critical step in the process of refining user value, but it’s also the easiest to overlook.
在本課中,我們將討論如何將使用者訪談進行回顧和綜合,轉化為清晰且有價值的見解。這是精煉使用者價值過程中的關鍵步驟,但也是最容易被忽視的。
Even if you’ve done all the previous steps correctly, you can still stumble at the final hurdle. Unless you synthesize those interviews into clear takeaways that you can use to refine the user value, all the time you have spent so far will have been wasted.
即使你已經正確完成了所有先前的步驟,你仍然可能在最後的關卡絆倒。除非你將那些訪談綜合成明確的結論,從而用來完善使用者價值,否則你迄今為止所花的所有時間都將白費。
There are three types of problems PMs face when they don’t synthesize their interviews into clear takeaways. First, they might either validate or invalidate hypotheses based on a small subset of memorable answers, rather than analyzing all interviews collectively. Let’s look at an example from Blue Apron, a meal kit delivery app. Blue Apron used to require at least a week’s notice to skip an upcoming order; otherwise, the user would automatically be billed for the following week’s meal kit. During user interviews, one user spent a long time emphasizing their belief that there should be a longer cancellation window available. Their strong feelings might bias the Blue Apron PM toward prioritizing this. However, if they were the only user bringing it up, it might not be worth pursuing.
當產品經理未能將訪談內容綜合成明確的結論時,會面臨三種類型的問題。首先,他們可能會基於少數令人印象深刻的回答來驗證或否定假設,而不是綜合分析所有訪談。我們來看看一個來自 Blue Apron 的例子,這是一個餐點套裝配送應用程式。Blue Apron 過去要求至少提前一週通知才能跳過即將到來的訂單,否則用戶將自動被收取下一週餐點套裝的費用。在用戶訪談中,有一位用戶花了很長時間強調他們認為應該提供更長的取消期限。他們的強烈意見可能會使 Blue Apron 的產品經理傾向於優先考慮這一點。然而,如果只有這位用戶提出這個問題,可能就不值得追求。
The second problem is that they might generalize learnings that only apply to a subset of the population. For instance, Grubhub might have heard 3 users say that their service is too expensive. Based on that, they might tell the pricing team that they should reduce prices. But if they debriefed and synthesized the information correctly, they would see that only users under 25 and living in college towns felt this way. This means there's an opportunity to provide differential pricing for this audience segment, or even define them as an "anti-audience" and focus on working professionals over 25 instead.
第二個問題是,他們可能會將只適用於某一部分人群的學習結果進行泛化。例如,Grubhub 可能聽到三位用戶說他們的服務太貴。基於此,他們可能會告訴定價團隊應該降低價格。但如果他們正確地進行了簡報和綜合信息,他們會發現只有 25 歲以下且住在大學城的用戶有這種感覺。這意味著有機會為這個受眾群體提供差異化定價,或者甚至將他們定義為「反受眾」,而專注於 25 歲以上的職業人士。
The third problem is that they might miss patterns that would allow them to reach deeper insights. For example, a PM at Peloton might interview 2 different audiences: users who have purchased their hardware, and users that only use the app. If users across both audiences said they wanted strong integration features with fitness tracking applications like Strava or Fitbit, then it’s the PM’s job to take a step back and identify this as a common theme regardless of audience segment. A PM who doesn’t do this synthesis may miss the opportunity to incorporate this learning into their design and development work.
第三個問題是,他們可能會錯過能夠讓他們達到更深入見解的模式。例如,Peloton 的產品經理可能會訪問兩個不同的受眾:購買了他們硬體的用戶,以及僅使用應用程式的用戶。如果這兩類用戶都表示希望與像 Strava 或 Fitbit 這樣的健身追蹤應用程式有強大的整合功能,那麼產品經理的工作就是退一步,將這視為一個不分受眾群體的共同主題。一位沒有進行這種綜合分析的產品經理可能會錯失將這些學習融入設計和開發工作的機會。
This is why it’s important to invest time in translating raw insights into a synthesized user value map.
這就是為什麼投入時間將原始見解轉化為綜合的用戶價值圖是如此重要。
There are two phases to this process: debriefing and synthesis. Debriefing happens immediately after each interview, which allows you to identify the most important insights while they are still fresh in your mind. It also provides an opportunity for you to switch up your approach to the next interview based on what you have learned so far. Synthesizing happens at the end of the entire interview campaign. This step allows you to look across all your interviews, identify patterns, and answer the initial questions you had from your hypotheses when you started the exercise. Let’s break down each of these processes, starting with debriefing.
這個過程分為兩個階段:回顧和綜合。回顧是在每次訪談後立即進行,這讓你能夠在記憶猶新的時候識別出最重要的見解。這也提供了一個機會,讓你根據目前所學調整下一次訪談的方法。綜合則是在整個訪談活動結束時進行。這一步驟讓你能夠審視所有訪談,識別模式,並回答在開始這項練習時從假設中提出的初始問題。讓我們從回顧開始,逐步解析這些過程。
After every interview, you should spend a few minutes writing down what you just heard. This allows you to begin identifying new insights, as well as incorporate lessons learned from the previous interview into the next one, ensuring each interview is more valuable than the last.
在每次訪談結束後,你應該花幾分鐘寫下你剛剛聽到的內容。這樣可以讓你開始識別新的見解,並將從前一次訪談中學到的經驗融入到下一次訪談中,確保每次訪談都比上一次更有價值。
Debriefing has four steps: document the user profile, document observations from the interview, extrapolate insights based on observations, and evaluate the interview process.
簡報有四個步驟:記錄用戶資料、記錄訪談觀察、根據觀察推斷見解,以及評估訪談過程。
You can reference the “Debrief” tab of the Refine User Value Template to follow along as we walk through this process.
您可以參考「Refine User Value Template」中的「Debrief」標籤,隨著我們一起進行這個過程。
Let’s start with documenting the user profile. It’s important to verify who you have spoken to and what attributes they have from your target audience. This will help you group responses based on user attributes, which will help you identify patterns later on.
讓我們從記錄使用者資料開始。確認您與誰交談過以及他們具備哪些目標受眾的屬性是很重要的。這將幫助您根據使用者屬性來分組回應,從而在後續識別模式時提供幫助。
Next, you’ll document observations from each interview. These observations should be about the user problem itself, and how the user experiences that problem. Above all, you want to make sure you understand what problem each user faced.
接下來,您將記錄每次訪談中的觀察結果。這些觀察應該圍繞使用者本身的問題,以及使用者如何體驗該問題。最重要的是,您要確保了解每位使用者面臨的問題是什麼。
This means asking, “If the problem was different, how was it different? What alternatives did they use to solve the problem? Why did they use those alternatives? What goals are they trying to achieve by solving the problem?”
這意味著要問:「如果問題不同,那是如何不同的?他們使用了哪些替代方案來解決問題?為什麼他們會選擇那些替代方案?他們試圖通過解決問題達成什麼目標?」
Answering these four questions allows you to verify if the observation either supports the hypothesis, contradicts the hypothesis, or introduces a new hypothesis.
回答這四個問題可以讓你確認觀察結果是支持假設、反駁假設,還是引入新的假設。
Next, you’ll want to extrapolate insights based on your observations. The goal of this step is to translate the surface-level observations from the last step into something more meaningful and specific to your project and its objectives. You can translate observations into insights for 3 categories: problem takeaways, severity takeaways, and alternative takeaways.
接下來,您需要根據觀察結果推導出見解。此步驟的目標是將上一階段的表面觀察轉化為對您的專案及其目標更有意義且具體的內容。您可以將觀察轉化為三個類別的見解:問題要點、嚴重性要點和替代方案要點。
After you’ve debriefed the content of the interview, you should evaluate the interview process itself and determine how to proceed. This helps you improve your interviews so that you get deeper and more meaningful insights each time.
在你完成訪談內容的回顧後,應該評估整個訪談過程並決定接下來的步驟。這有助於你改進訪談,使每次都能獲得更深入且有意義的見解。
If you don’t take the time to evaluate your interview process and adjust if necessary, you’re likely to make the same mistakes, face similar challenges across interviews, and encounter diminishing returns.
如果你不花時間評估你的面試過程並在必要時進行調整,你很可能會犯下相同的錯誤,在面試中面臨類似的挑戰,並遭遇收益遞減的情況。
After completing an interview you can decide to do one of three things. One, don’t change anything. This normally happens when you are early on in your research and haven’t yet reached saturation on your questions. You don’t want to change anything right away if things are going well, because there is likely more to be learned from your existing plan. Two, you can adjust your questions. This can happen when you’ve reached saturation on a particular question, meaning you’re hearing the same thing over and over again from users and aren’t learning anything new. Three, you can stop interviewing. Maybe you’re hearing the same things consistently, and you don’t think you’re going to hear anything remarkably different or gain new insight from continuing to have additional conversations.
完成訪談後,你可以選擇做三件事之一。第一,不做任何改變。這通常發生在研究初期,當你對問題的飽和度尚未達到時。如果一切進展順利,你不想立即改變任何事情,因為現有計劃中可能還有更多可以學習的內容。第二,調整你的問題。這可能發生在你對某個問題達到飽和時,意味著你從用戶那裡聽到的都是重複的內容,沒有學到任何新東西。第三,停止訪談。也許你聽到的內容始終如一,並且認為繼續進行更多對話不會聽到任何顯著不同的內容或獲得新的見解。
For example, if you’ve heard from 4 participants that they have the exact same concern about something, participants 5 and 6 aren’t likely to yield new insight.
例如,如果你從四位參與者那裡聽到他們對某件事情有完全相同的擔憂,那麼第五和第六位參與者不太可能提供新的見解。
Let’s go through a debrief exercise using some illustrative responses from users that the PM at Gusto might have interviewed. Check out the visual here if you need a reminder of the initial user value hypothesis the PM defined with their manager.
讓我們透過一些來自 Gusto 的產品經理可能訪談過的用戶的示例回應,進行一次簡報練習。如果你需要回顧產品經理與其經理共同定義的初始用戶價值假設,請查看這裡的視覺資料。
Pause here and review these interview notes from a conversation with Brandy, a CPO at a growing startup. Then, answer these three questions: How do these responses compare to the initial problem hypothesis? What might you extrapolate from these responses? How, if at all, would you change your interview approach?
請在此暫停,並查看這些來自與一位成長中初創公司首席產品官 Brandy 的對話訪談筆記。然後,回答這三個問題:這些回應與最初的問題假設相比如何?你可以從這些回應中推斷出什麼?如果有的話,你會如何改變你的訪談方法?
Let’s start by answering the first question. Brandy’s responses indicate that Gusto was right to hypothesize that growing businesses have a ton of setup to do when onboarding new employees, and that it can be easy to make mistakes. The team can also extrapolate a few key takeaways from Brandy’s responses. A problem takeaway is that the mistakes made include data entry errors as well as missed deadlines—a problem the team hadn’t previously highlighted in the user value map. In terms of severity, Brandy’s experience of missing a deadline, which resulted in an employee not having benefits for the first month of work, shows us that this type of mistake can be particularly severe. This might indicate a high willingness to pay for a robust solution. In evaluating their interview approach, the team might choose to probe for deadline tracking in future interviews to understand how big of a problem this is, and for whom. This will enable them to determine whether or not this is an issue worth prioritizing in feature design.
讓我們先回答第一個問題。Brandy 的回應顯示,Gusto 的假設是正確的,即成長中的企業在為新員工入職時有大量的設置工作要做,而且很容易出錯。團隊還可以從 Brandy 的回應中推斷出幾個關鍵要點。一個問題要點是,所犯的錯誤包括數據輸入錯誤以及錯過截止日期——這是團隊之前在用戶價值圖中未曾強調的問題。就嚴重性而言,Brandy 錯過截止日期的經驗,導致一名員工在工作的第一個月沒有福利,這告訴我們這類錯誤可能特別嚴重。這可能表明用戶願意為一個強大的解決方案支付高價。在評估他們的訪談方法時,團隊可能會選擇在未來的訪談中探討截止日期追蹤,以了解這個問題有多大,以及對誰影響最大。這將使他們能夠確定這個問題是否值得在功能設計中優先考慮。
Now, let’s answer the same three questions using an interview Gusto conducted with Joseph, the founder and CEO of a seed-stage startup. How do these responses compare to the initial problem hypothesis? What might you extrapolate from these responses? How, if at all, would you change your interview approach?
現在,讓我們使用 Gusto 與一位種子階段初創公司的創辦人兼 CEO Joseph 進行的訪談來回答相同的三個問題。這些回應與最初的問題假設有何不同?你可以從這些回應中推斷出什麼?如果有的話,你會如何改變你的訪談方式?
Let’s start with observations. Joseph’s interview surfaced a new hypothesis: storing sensitive employee data is painful, and an important aspect of an onboarding tool. We can extrapolate problem and severity takeaways from these observations. The problem may be larger in scope than the team initially anticipated, including not only data entry but also secure data storage and work authorization verification. Joseph’s work authorization mistake in particular indicates that the consequences of these onboarding mistakes may be more severe than originally expected. This is a high-severity problem. In terms of evaluating the interview approach, one adjustment the team would likely make is to start asking questions about securely storing employee data and managing work authorization after hearing about these problems from Joseph.
讓我們從觀察開始。Joseph 的訪談提出了一個新的假設:儲存敏感的員工資料是痛苦的,並且是入職工具的一個重要方面。我們可以從這些觀察中推斷出問題和嚴重性。問題的範圍可能比團隊最初預期的更大,不僅包括資料輸入,還包括安全的資料儲存和工作授權驗證。特別是 Joseph 的工作授權錯誤表明,這些入職錯誤的後果可能比原先預期的更為嚴重。這是一個高嚴重性的問題。在評估訪談方法方面,團隊可能會做出的調整是,在聽到 Joseph 提到這些問題後,開始詢問有關安全儲存員工資料和管理工作授權的問題。
After you’ve completed and debriefed all of your interviews, you need to synthesize your learnings across all interviews to effectively answer the questions you set out to answer.
在完成並回顧所有訪談後,您需要綜合所有訪談中的學習成果,以有效回答您所設定的問題。
If you don’t synthesize across interviews, you could end up generalizing anecdotal evidence from your most vocal participants, as opposed to identifying trends and patterns. You could also find yourself cherry-picking responses to fit your narrative or hypotheses, as opposed to learning new things from your users even if they disprove what you believed initially.
如果你不綜合分析各次訪談,你可能會將最健談的參與者的軼事證據泛化,而不是識別趨勢和模式。你也可能會挑選符合自己敘述或假設的回應,而不是從用戶那裡學到新事物,即使這些新事物推翻了你最初的信念。
We’ll introduce a five-step process to synthesize your learnings. To guide us through this process, we’ll refer to the “Synthesis” tab of the Refine User Value Template.
我們將介紹一個五步驟的過程來綜合你的學習成果。為了引導我們完成這個過程,我們將參考「精煉用戶價值模板」中的「綜合」標籤。
The first step is clustering based on the problems faced by users.
第一步是根據使用者面臨的問題進行分組。
This helps you see the frequency of different problems that users faced, and whether those problems were similar or different to those you hypothesized. Generally, you want to take a closer look at the problems most frequently mentioned by users, and determine the root of those problems. This helps you add additional nuance to the problem description and the reasons behind the problem.
這有助於你了解使用者面臨的不同問題的頻率,以及這些問題是否與你假設的相似或不同。通常,你會想要仔細觀察使用者最常提到的問題,並找出這些問題的根本原因。這有助於你為問題描述和問題背後的原因增添更多細微差別。
After you have identified the problems to focus on, you can look for patterns across user profiles.
在確定要專注的問題後,您可以在用戶資料中尋找模式。
This helps you understand if only certain user profiles experience a problem or if all groups of users experience a similar problem.
這有助於你了解是否只有某些使用者群體遇到問題,或是所有使用者群體都遇到類似的問題。
You can ask yourself questions like, “Did different user profiles experience different problems or have different goals? Do different user profiles experience the problem at different frequencies? Do different user profiles experience the problem at different levels of severity?”
你可以問自己這樣的問題:「不同的使用者角色是否遇到不同的問題或有不同的目標?不同的使用者角色是否以不同的頻率遇到問題?不同的使用者角色是否在不同的嚴重程度上遇到問題?」
After you have established user problem combinations, you can identify patterns across alternatives and severity.
在確立使用者問題組合後,您可以識別不同選擇和嚴重程度之間的模式。
This helps you understand which problems are more severe across user groups, and how well the alternatives they use are working.
這有助於你了解哪些問題在用戶群中更為嚴重,以及他們所使用的替代方案效果如何。
You can ask yourself questions such as, “Are some problems more severe than others across different user profiles? What causes the difference (or delta) in severity? Do some problems have more or better alternatives?”
你可以問自己一些問題,例如:「在不同的使用者檔案中,是否有些問題比其他問題更嚴重?是什麼原因導致嚴重程度的差異?有些問題是否有更多或更好的替代方案?」
Let’s return to our Gusto example to practice clustering users and identifying patterns based on the illustrative responses we shared. In Joseph and Brandy’s interviews, we heard about data entry mistakes, missed timelines, and data storage security concerns. Brandy and Joseph represent two different user profiles: Brandy already has HR systems, while Joseph does not.
讓我們回到 Gusto 的例子,練習將用戶進行分群並根據我們分享的示例回應識別模式。在 Joseph 和 Brandy 的訪談中,我們聽到了數據輸入錯誤、錯過的時間表以及數據存儲安全問題。Brandy 和 Joseph 代表了兩種不同的用戶形象:Brandy 已經擁有 HR 系統,而 Joseph 則沒有。
These problems show up in different ways for these two types of users. For those with existing HR management solutions like Brandy, the most severe problems are data entry and missed timelines. Secure data storage isn’t as much of a challenge because the existing tools provide this functionality. Meanwhile, users like Joseph who don’t yet have HR solutions struggle primarily with missed timelines and secure data storage. These problems are high severity, and the existing manual approach doesn’t offer solutions to these challenges.
這些問題在這兩類用戶中以不同的方式顯現。對於像 Brandy 這樣已有 HR 管理解決方案的用戶來說,最嚴重的問題是數據輸入和錯過時間表。由於現有工具提供了這項功能,安全的數據存儲並不是太大的挑戰。與此同時,像 Joseph 這樣尚未擁有 HR 解決方案的用戶主要面臨的是錯過時間表和安全數據存儲的困擾。這些問題的嚴重性很高,而現有的手動方法無法提供這些挑戰的解決方案。
Now that you’ve identified key problem clusters, profiles, and patterns, the next step is to complete your user value map. This map should capture the refined hypotheses for how users will get value from your feature. Start with the hypotheses you gathered in your manager briefing, and make updates to each based on the insights you extrapolated from your user interviews.
現在你已經識別出關鍵的問題群組、用戶特徵和模式,下一步就是完成你的用戶價值地圖。這張地圖應該捕捉到用戶從你的功能中獲得價值的精煉假設。從你在經理簡報中收集的假設開始,並根據從用戶訪談中推斷出的見解對每個假設進行更新。
Let’s complete the user value map for Gusto. As a reminder, here are the user profile and user problem hypotheses we landed on in our manager briefing. Pause the video here. What changes would you make to these hypotheses based on your insights from user interviews? What additional detail would you add to help your team hone in on a solution?
讓我們完成 Gusto 的用戶價值地圖。提醒一下,這是我們在經理簡報中確定的用戶檔案和用戶問題假設。請在此暫停影片。根據您從用戶訪談中獲得的見解,您會對這些假設做出哪些更改?您會添加哪些額外的細節來幫助您的團隊專注於解決方案?
Let’s start by refining our user profile and user problem hypotheses. The biggest change we need to account for here is adding in the user problem around secure data storage. We can also add some more information about the severity of this problem, and the alternatives that exist today. The updated map might say: “Through conversations with potential customers like Joseph, the team realized that this feature could help both current Gusto users and non-users. The team was also able to hone in on specific individuals that would benefit most from this feature—namely, founders and other executives who are growing their teams.”
讓我們從完善用戶檔案和用戶問題假設開始。我們需要考慮的最大變化是加入有關安全數據存儲的用戶問題。我們還可以添加更多有關此問題嚴重性的信息,以及當今存在的替代方案。更新後的地圖可能會說:「通過與像 Joseph 這樣的潛在客戶的對話,團隊意識到這個功能可以幫助現有的 Gusto 用戶和非用戶。團隊還能夠專注於那些最能從這個功能中受益的特定個人——即創始人和其他正在擴展團隊的高管。」
For user problem, the team validated the pain point around making mistakes, and got more specific about the most common types of mistakes: data entry errors and missed deadlines. They also identified a new problem: concerns around personal data storage. Across all of these problems and multiple user types, the team found that severity was high, and existing alternatives were limited. After refining the user profile and user problem hypotheses, the team decided to slightly update their user goal: instead of just focusing on efficiency, the feature should create a secure and efficient onboarding experience.
針對使用者問題,團隊驗證了關於犯錯的痛點,並更具體地指出最常見的錯誤類型:資料輸入錯誤和錯過截止日期。他們還識別出一個新問題:對個人資料存儲的擔憂。在所有這些問題和多種使用者類型中,團隊發現嚴重性很高,現有的替代方案有限。在精煉使用者檔案和使用者問題假設後,團隊決定稍微更新他們的使用者目標:功能不僅要專注於效率,還應創造一個安全且高效的入職體驗。
The final step of synthesis is to reevaluate the project based on your conclusions.
綜合的最後一步是根據你的結論重新評估專案。
If the conclusions you reached in the last step support or add to your user value hypotheses, then you should continue exploring the problem and move on to refining the business value. However, if the conclusions contradict your hypotheses or disprove key assumptions, you should reevaluate if this is the right problem to be working on, or if other problem hypotheses might be more relevant and impactful for the business.
如果您在上一步得出的結論支持或增強了您的用戶價值假設,那麼您應該繼續探索該問題,並進一步完善商業價值。然而,如果結論與您的假設相矛盾或推翻了關鍵假設,您應重新評估這是否是值得投入的正確問題,或者其他問題假設是否對業務更具相關性和影響力。
If this is the case, you can bring a few options to your manager: you can conduct another short round of interviews to prove a new hypothesis, you can stop working on the problem altogether and move to a different project, or you can proceed with a refined view of a new problem. When bringing these options to your manager, you should have a recommendation on how you want to proceed and why you want to proceed that way.
如果是這樣的情況,你可以向經理提出幾個選擇:你可以進行另一輪短期訪談來證明新的假設,你可以完全停止處理這個問題並轉向其他專案,或者你可以以新的問題的精緻觀點繼續進行。在向經理提出這些選擇時,你應該有一個建議,說明你想如何進行以及為什麼要這樣進行。
For the Gusto example, the conclusions reached indicated that this feature has the opportunity to deliver real value to the target user. Since their user value hypotheses were largely supported, all signs point to continuing the project.
以 Gusto 為例,得出的結論顯示此功能有機會為目標用戶帶來真正的價值。由於他們的用戶價值假設大多獲得支持,所有跡象都指向繼續推進該項目。
This completes our interview process, which helped us validate our initial user value hypothesis. In the next section, we’ll explore how to validate our business value hypotheses.
這完成了我們的訪談過程,這幫助我們驗證了最初的用戶價值假設。在下一部分,我們將探討如何驗證我們的商業價值假設。
- Turning raw interview data into valuable insights requires 2 steps: debriefing after each interview, and synthesizing across the whole interview campaign.
將原始訪談數據轉化為有價值的見解需要兩個步驟:在每次訪談後進行簡報,並綜合整個訪談活動的結果。 - In the post-interview debrief, document the user profile and early observations, and then extrapolate takeaways about the problem, severity, and alternatives.
在訪談後的簡報中,記錄用戶資料和早期觀察,然後推斷出關於問題、嚴重性和替代方案的要點。 - To synthesize across interviews, follow a 5-step process: cluster users by problem, identify patterns across user profiles, identify patterns across alternatives and severity, create a user value map, and finally reevaluate the feature opportunity based on your conclusions.
要綜合各次訪談,請遵循五步驟流程:按問題將用戶分組、識別用戶資料中的模式、識別替代方案和嚴重性中的模式、創建用戶價值圖,最後根據結論重新評估功能機會。
You've completed this lesson. How useful was it?
Great work! You've completed this lesson. How useful was it for you?
做得好!你已完成這堂課。這對你有多大幫助呢?